- Email: [email protected]
Design and implementation of Intelligent transplanting system based on photoelectric sensor and PLC
Accepted Manuscript Design and implementation of Intelligent transplanting system based on photoelectric sensor and PLC Jin Xin, Zhao Kaixuan, Ji Jiangtao, Du Xinwu, Ma Hao, Qiu Zhaomei
PII: DOI: Reference:
S0167-739X(18)30634-4 https://doi.org/10.1016/j.future.2018.05.034 FUTURE 4209
To appear in:
Future Generation Computer Systems
Received date : 22 March 2018 Revised date : 23 April 2018 Accepted date : 15 May 2018 Please cite this article as: J. Xin, Z. Kaixuan, J. Jiangtao, D. Xinwu, M. Hao, Q. Zhaomei, Design and implementation of Intelligent transplanting system based on photoelectric sensor and PLC, Future Generation Computer Systems (2018), https://doi.org/10.1016/j.future.2018.05.034 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Design and implementation of Intelligent transplanting system based on photoelectric sensor and PLC Jin Xin1,2 Zhao Kaixuan1 Ji Jiangtao1,2* Du Xinwu1 Ma Hao1
Qiu Zhaomei1
1. College of Agricultural Equipment Engineering, Henan University of Science and Technology, Luoyang 471003, China; 2. Collaborative Innovation Center of Machinery Equipment Advanced Manufacturing of Henan Province, Luoyang 471003, China. * Corresponding author: Ji Jiangtao, Professor, dedicated to the research on new technology and theory of agricultural machinery and intelligent agricultural equipment. Address: 48 Xiyuan Road, Luoyang 471000, Henan Province, China. Tel:+86-379-64877837, Email: [email protected] Abstract: Industrial seedling rearing is an important part of seedling transplanting technique. At present, there exists the phenomenon that soil base on the tray may have empty seedling on it, which will lead to a large number of leakage and reduce the yield in the course of automatic transplanting in the field. To solve this problem, automatic seedling identification method is studied to skip picking empty soil bases. An intelligent transplanting system are designed with the use of the picking mechanism five-bar and fixed-axis gear train, the seedling tray conveying mechanism (with transverse and longitudinal seedlings feeding function), the eccentric disc parallel four-bar duck mouthpiece planting mechanism, the electric sensor for seedling detection and identification of the seedling tray, the position sensor, stepper motor and PLC control system. By use of computer logic programming and control, it can be used to identify effectively whether the soil base is empty or not, and to control the automatic transmission of seedling tray. The tray is moved to quickly skip the action path of the manipulator to avoid the picking of empty soil base. The system was tested by using pepper seedlings on indoor condition. The results showed that when the transplanting frequency was 90 / min, the rate of success seedling picking was 88.23% and the leakage rate was 16.46%, which could meet the requirement of pepper transplanting. Compared with the transplanting model without seedling identification, the average rate of leakage was reduced about 12%. The research provides a useful reference for the development of intelligent agricultural transplanting technology and equipment. Keywords:
1
Intelligent agricultural; Transplanting system;
Automatic recognition;
PLC control;
Logic programming
Introduction
The vacancy of seedling in factory seedling cultivation is an important factor leading to the leakage of subsequent mechanical transplanting. In the whole automatic transplanting process, the number of seedlings planted in the unit area will be reduced if the vacancy of seedling cannot be effectively identified and avoided which will reduce the production. At present, the main way to deal with this problem is to fill the seedlings after planting. The whole process is characterized by high labor intensity, low efficiency and high cost [1-3]. Therefore, it is urgent to improve the intelligent level of mechanical transplanting technology. The advanced sensor and computer logic control technology are applied to the transplanting equipment to realize the independent identification and selective operation of the machine. Many useful researches have been done in the field of seedling identification and intelligent transplanting. For example, Tai et al. [4] used machine vision for automatic seedling raising in greenhouse, a recognition system is developed, which can automatically identify the empty pot on the seedling tray. Ryu et al. [5] designed an automatic transplanter system including machine vision, which can realize the functions of automatic conveying, grabbing and planting seedling. Humphries S et al. [6] used the image tracking algorithm to divide the seedling plants into different parts with similar geometric characteristics, and realize the classification and recognition of the leaves, stems and main *
Corresponding author: Ji Jiangtao, Professor, dedicated to the research on new technology and theory of agricultural machinery and intelligent agricultural equipment. Address: 48 Xiyuan Road, Luoyang 471000, Henan Province, China. Tel:+86-379-64877837, Email: [email protected]
stems of the seedlings. Mizuochi et al. [7] used machine vision technology to develop the information recognition system of the seedling, according to the shape characteristics of the cotyledon leaf circumference of the seedling, to identify the healthy seedling, and to eliminate the diseased and abnormal seedlings. Albertus et al. [8] designed the multi-disc automatic conveying device, and based on the vision recognition system, realized the automatic seeding supplying and seeding feeding; Hu Fei et al. [9] designed a machine vision system for transplanting seedlings, which was used to measure the leaf area of seedlings in each pot in real time, and to determine the suitable grasping position for transplanting seedlings. Hou et al. [10] studied the identification technology of healthy seedling, and established a comprehensive evaluation method based on multi image feature of healthy seedling, which can realize the identification of robust seedling and the location of rootstock. Yang et al. [11] based on monocular vision technology, designed an algorithm to obtain the information of transplanting suitability of potted seedlings, and used the Harris corner detection algorithm to obtain the information of main stem of potted seedlings. He et al. [12] studied the method of optimizing the path of the seedling supplying, and used machine vision to obtain the information of the seedlings, to control the executer to grab the healthy seedlings to complete the transplanting. In order to realize the precise positioning control of transplanting and seeding device, the adaptive fuzzy PID controller and its control rules were designed by Liu et al. [13] However, most of the above studies are focused on the process of seedling raising in factories. The health status identification, pot location and culling control technology of seedling leakage are seldom reported, and few reports have been made on the control technology of the transplanter to identify and evade the seedling vacancy in the pot tray. In this paper, in order to solve the problem of self-identification and avoiding the vacancy of seedling in transplanter, this paper analyzes and studies the seedling identification and positioning control of automatic transplanting machine, and intelligent transplanting system is developed with mechanical design, the sensor, computer programming and PLC control, etc. The system recognizes the vacancy of seedling by photoelectric sensor, localizes the position of the pot with seedlings vacancy by combining the position sensor, and transmits the position information to the PLC controller to guide the seedling feeding mechanism to evade that pot, to achieve the goal of precise seedling extraction by the seedling manipulator. Using photoelectric sensor, position sensor and computer logic programming to replace machine vision to obtain the information of seedling vacancy, which avoids the interference of the machine vision caused by the complex working environment of transplanting, and can provide a reference for the intelligent design of transplanting equipment.
2
Structure and working principle of transplanting device
2.1
Transplanting object
The transplanting object is the pepper seedlings, and the variety is Hejiao 12, which originated from Mexico. The seedling plate is a conical tray of 128 pots. The depth of the pot is 41.5mm. The spacing between pots is 36.5mm. The of size of the pot top is 30.5mm × 30.5mm. The size of the pot bottom is 15mm × 15mm, and the volume of the cavern is about 21ml. The seedling substrate was mainly peat and vermiculite with a volume ratio of 3: 1; The way of seedling raising is to adopt factory cultivation, that is, precision seeding equipment is used to complete filling the pot, seeding, covering and compacting in one operation[14-18], then replenishing water regularly, controlling temperature and light, etc. The seedlings at age of 35 days are suitable for transplanting. At this time the seedlings grow 5~6 leaf and root has matrix winding package, as shown in Figure 1a; The industrial seedling breeding is based on the idea of precision seeding, that is, one seed for one pot. Although the environment, water, fertilizer and other crop growth conditions of seedling raising process are standardized, but because of the difference of seed quality and artificial control growth conditions, there is no guarantee that every seed can emerge. At present, the seedling emergence rate of Chinese factory is over 90%, as shown in Figure 1b. The average diameter of the seedling stem was 3.56mm (about 10mm from the top surface of the seedling pot), the height of the seedling stem was 40mm ~ 51mm, and the moisture content was about 52%.
a
Fig g.1
b
Seedlings at age of 35 daays
a Single S potted seeedling.
d, Diaameter of stem; h, Height of steem; l, Seedling leaf broadening
b Seedling tray with vacan ncy
2.2 2 Structuree of transplan nting device The transpplanting devicce is mainly composed c off seedling tray y conveying mechanism, m aautomatic seeedling collectting meechanism andd rotary plannting mechaanism, as shoown in Figu ure 2. The design d of auutomatic seedling collectting meechanism refeers to the com mbined plantin ng-taking meechanism of five-bar f and fixed f axis geaar train [19]. Itt mainly conssists of gear transmiission box (a pair of straig ght cylindric al gears of th he same size meshing), ddouble crank, connecting rod, r seeedling arm, seeedling claw (a pair of seeedling needlee, push rod and a ring) and other parts ((as shown in Figure 2b). The T two o crank are ddriven by gearr to rotate in the opposite direction, and the power is i transferredd from the con nnecting rod and according to the predeterm thee seedling arm m to the seedlling claw, so that the seeddling claw is reciprocated r mined track, and thee seedlings arre taken and dropped d in th he fixed positiion. The desiign requiremeent of seedlinng tray conveying mechannism is that t each pott on the tray can be moved intermitteently to the position p to bee taken accorrding to a cerrtain order. This T pro ocess avoids interference between seedling tray annd seedling claw, c and sho ould be conveenient to recover empty tray t after taking seeedling. The design d adopts two parallell synchronouss chain drivees to realize tthe longitudin nal tray feediing. Th he guide rail llead screw mechanism waas used to carr rry the seedlin ng tray transv versely. Theyy are all driveen by step mootor, and d the distance to move is equal to the distance bettween pots, which w is 36.5 mm (see Figgure 2a). Thee rotary plantting meechanism adoopts duckbillled planting mechanism with eccenttric disc and d parallel fouur-bar [20], which w is maiinly com mposed of foour duckbill planters, cam m sliders, ecccentric diskss, left and rig ght plates annd eccentric disc slides. The T meechanism cann realize rotaryy planting an nd keep the duuckbill planteer always verttical downwaard.
Fig g.2
Plant tran nsplant device for pepper seeedling
a Tray T conveying mechanism b Combined C Seed dling collecting Mechanism of Five-bar and fiixed Axial Wheeel Train. 1, electric e control cabinet; 2, boddy frame; 3, ducckbilled plantingg mechanism with w eccentric diisc and parallel four-bar; 4, converter motor 5, seedling s claw; 66, Seedling arm m; 7, double cran nk; 8, connectinng rod; 9, driven gear; 10, Two o parallel synchhronous driven chain; 11, Seedling tray; 12, Guide rail screw s mechanissm; 13, transverrse moving step p motor
2.3 3 Working principle of transplantin ng device The autom matic control technology is used to connnect the mottion of seedliing conveyinng, picking an nd planting. The T wo orking principple block diaggram of the trransplanting ddevice for pep pper seedling gs is shown ass shown in Fiigure 3. a Move the tray y
Stop at the ppicking
horizontally
positionn
Finish F serving seedlings in row horizontally h
Yes Move e the tray once to the picking
No
position
seedlingg tray conveying mecchanism
b
Autom matic seedling pickingg mechanism Picking the t seedling
c
Pla anting mechanism
Burrow and drrop
by y claw
Conveeying the
Carry seedlin ng
seeedling
Receive seedling by Releasee and drop
the planter
Fig.3 Working principle of transplanting device a Seedling tray conveying step b Automatic seedling taking c Seedling planting step
The seedling taking mechanism and the rotary planting mechanism always do reciprocating motion according to a certain path, and the power is provided by frequency conversion motor and driven chain. The transmission ratio of two mechanisms is set to 1: 4, that is, each time the seedling claw reaches the lowest point (the position of sapling), the corresponding duckbill transplanter rises to the highest point (seedling position) and matches it. The transverse and longitudinal feeding of the seedling tray by conveyer mechanism is powered by stepping motor. When the seedling is taken out completely, the step motor drives the seedling tray to move to the next pot transversely. The claw moves downward with the seedling, and when moves to the dropping position, the duckbilled planter rises to the position where the seedling is received, and the seedling claw releases the seedling and place it into the planter. The planter then moved down until the lowest point was completely opened, and implanted the seedling into the field. At the same time, the planter continued to grab the next seedling which had been moved to the position for picking up [21-24]. In this way, when the seedling grabs the last seedling of a row, the step motor drives the tray to move to the next pot longitudinally, so that the next row of pots on the edge of the tray is moved to the position to be picked up, and then continue to complete the feeding, taking, and planting. The whole process moves in cycles, until the whole tray of potted seedling were transplanted.
3
Design of vacancy recognition scheme for seedling
3.1 Requirements for detection and identification of seedlings The intelligent transplanting of pepper seedling requires that it should be able to identify whether there is a vacancy of seedlings or not in the seedling collection stage, and to grasp the seedlings, and move quickly to the next seedling position if there is no seedling. The following conditions should be satisfied in order to realize the rapid detection and identification of the seedlings: 1. After the seedlings were raised in the pot, the top branches and leaves of the seedlings were mostly cross-occluded, which make it difficult to distinguish the branches and the leaves of the seedlings [25,26]. This problem should be solved or avoided when the sensor detects whether there is a seedling or not. 2. The interference of adjacent seedlings should be overcome. 3. Not subject to restrictions on testing materials. 4. Sensor with high accuracy and sensitivity is adjustable. 5. Sensor detection and identification signal feedback ha high speed and is easy to use. 6. Sensor structure size is small; light is weight, and easy to install and adjust position. 3.2
Selection and verification of the sensor for seedling detection
(1) Sensor selection In order to meet the above conditions, on the basis of analyzing the researches on the detection and identification of nonmetallic bodies, combining with the structure and operation characteristics of the transplanting device, BGS-S08N background inhibitory diffuse reflectance photoelectric sensor is used to detect and identify the pepper seedlings [27-30]. The specific parameters are shown in Table 1. The photoelectric sensor integrates a transmitter and two receivers, with simple structure, small size and easy installation; At the same time, it has the function of background suppression, which can effectively avoid the interference of strong background light beyond the detection distance, which is more reliable than the conventional reflective sensor detection. The sensitivity can be adjusted, which make it easy to adjust the detection distance.
Table 1
Model
Dimensioons
Technical pparameters of th he photoelectricc sensor
deteection
Dettection
mode
disstance //mm
onse respo
voltag ge
current
Control
levels of
time /ms
/VDC C
/mA
output
protection
<30
NPN
IP67
Backg ground BGS-S08N N
supprression difffuse
100~80
<0.5 0
0 10~30 (±10% %)
refleection
(2) Feasibiility Test of Seedling S vacancy in pots In order too verify the feasibility off BGS-S08N background inhibitory diffuse reflecttance photoelectric sensorr to dettect and identtify the planttlets, the senssitivity of the detection an nd identificatiion of non-me metallic materiials was studied. Th he response tim me and the acccuracy of seedling stem rrecognition under natural light l were tessted. Test methood: the seedliing tray is fix xed, the photooelectric senssor is parallel to the tray suurface. The sensor s is movving at a certain speeed to detect thhe seedling sttem. The test coonditions weere as followss: 3 tray of ppepper seedliings were sellected and soome seedlings were removed ran ndomly to maake 40 vacanncy seedlingss; It was testted under sim mulated naturral light in thhe laboratory y. The transveerse disstance betweeen photoelecttric sensor an nd seedling sttem is about 35mm, and the vertical diistance is abo out 15mm to the surrface of seedlling bowl. Thhe sensors weere moved traansversely baack and forth at a speed off above 4 potts / s. After each e row w, the photoeelectric sensoor moves forw ward to the nnext row and continues to detect. The rresponse of the t sensor to the tray is shown inn Figure 4. The resultss showed thatt 320 pots in 40 rows, andd 40 vacancy seedlings weere detected inn the experim ment. The triggger reaaction speed of the sensoor is relatively fast (by thhe output sign nal indicator light). All oof pot with seedlings s cann be ideentified, and tthere are 2 vaacancy seedlin ngs failed to detect, and th he accuracy of o recognitionn of vacancy seedling s is 955%. The main rreasons for missing m detecttion are the innterference of adjacent seeedlings, suchh as the tilting g of the right and lefft seedlings, aand the lodgiing of the seedlings in thhe back row, which result in the seedliing leaves bllocking the liight beaams of the sensors and cauusing misjudg gment.
Fig g.4
Response Test of BGS-S S08N photoelecctric Sensor
Accordingg to the expeerimental results, it can bbe concluded d that it is feasible f and effective to use BGS-S008N pho otoelectric seensor to idenntify the seeedling stem. Then in thee transplantin ng device, foour BGS-S08 8N photoelecctric sen nsors are arraanged separaately (as show wn in Figuree 6) for autom matic identiffication of thhe whole seedling conveyying pro ocess. The seeedling will be b taken out when the pott with seedlin ng. Otherwise, the pot wiill be abando oned. The dessign ideea of movingg the tray to the next seeedling pot whhen vacancy is detected should have the same feasibility, andd its imp plementationn effect needs to be verified by experim ment. 3.3 3 Identificaation schemee of seedling vacancy durring seedling g tray convey ying The installlation height of the BGS--S08N photoeelectric senso or was adjustted to detect the seedling stem below the misjudgmentt caused by the leaaf (about 10m mm / 15mm m from the top surface oof the pot), which could d avoid the m
interleaving of the leaves. The T sensitivity y of the sensoor can be adjjusted to overrcome the intterference off the seedlings in thee back row. Figure 5 is a schematic diag gram of sensoor detection and a identificaation. It can bbe seen from the diagram that t no matter wherre the seedlinng position is taken, the beam emitted by the sen nsor must alw ways be perp pendicular to the seeedling stem, that is, paralllel to the su urface of the pot. At the same time, th he sensor shhould have a certain distaance relative to the sseedling stem m, to prevent the t interferennce of the seeedling in the back row andd the interferrence of the trray. In the process oof the seedling tray convey yor moving thhe seedling to o the seedling g position, thhe seedling steem goes throuugh thee beam of thee sensor, and the sensor is triggered to pprove that th here is a seedlling. If the seensor is not trriggered, therre is no seedling.
Fig g 5.
Schematiic Diagram of Detection D and recognition off Seedling
1, BGS-S08N B phootoelectric sensoor;
2, Seedling g tray;
3, Lonngitudinal shift driven chain; 4, Long pin shhaft
Because thhe mechanism m of seedling g supply neeeds to supply the seedling gs both transvversely and longitudinally l y, it neeeds the coopeeration of fouur photoelectrric sensors too complete th he recognition n of seedling vacancy [2, 13, 31]. The speccific arrrangement off the sensors inn the automatic transplantting device is shown in Fig gure 6.
Fig g. 6 Installation n of photoelecttric sensors 1, Stepping S motor for longitudinaal move; 2 (6 10 0 12), BGS-S088N photoelectriic sensor; 3, Meechanism for loongitudinal mov ve; 4, Mechanissm for transverse movve; 5, Seedling tray; 7, Guide rail r screw mechhanism; 8, Stepp ping motor for transverse movve; 9, Seedling claw c position sen nsor; 11, Seedlinng claw
It can be seen from Figure 6 that the implementation of seedling vacancy identification in the tray conveying process mainly depends on the transverse and longitudinal movement of step motor and four BGS-S08N photoelectric sensors. In the picture, the seedling tray is located at the leftmost end of the mechanism. The seedling claw is pointing to the pot No. 1. Sensor 12 and sensor 10 are on the left and right sides of the seedling claw, and placed in the right side of the longitudinal centerline of pot 2 and pot 1 which are fixed to the frame. The position of sensor 6 and sensor 2 are placed under the lateral centerline of pot 1 and pot 8 (fixed with the machine frame and moving with transverse movement). The realization process of seedling vacancy recognition is as follows: 1. The tray is fed longitudinally and the pot 1 is moved to the seedling position and the longitudinal motion motor stops. If the sensor 6 detects that there is a seedling in the pot cell 1 in the process of moving forward, then it waits for the seedling to be picked up, and enter the step 2 the once the seedling is taken out; If sensor 6 does not detect a seedling in pot cell 1, then entered step 2 immediately. 2. Move the tray to the next pot transversely, and move the pot 2 to the position to be taken. Wait for the seedling claw to take the seedling, and move the tray to the next right pot after taking out. In the process of right shift, if the sensor 12 does not detect the seedlings in the next pot, then immediately move the next pot to the position to be taken. 3. Repeat step 2 until the seedling in the pot 8 in the tray is taken out. 4. Move the tray to the next pot longitudinally, and move the pot 9 to the position to be taken, then stop the longitudinal motion motor. If the sensor 2 detects the presence of seedlings in pot 9 as it moves forward, it waits for the claw to taking the seedlings, and enter step 5 after the seedlings are taken out. If sensor 2 does not detect seedlings in cell 9, then enter step 5 immediately. 5. Move the tray to the next left pot, and move pot 10 to the position for taking out. Wait for the claw to pick up seedling, and move the tray to the next left pot. During the left shift, if the sensor 10 does not detect a seedling in the pot, then move to next left pot immediately, and move the next pot to the position to be taken. 6. Repeat step 5 until the seedling in pot 16 is taken out. 7. Move forward longitudinally, and move pot 17 to the position to be take, and then repeat step 2 until the whole seedlings are taken out.
4
Design of intelligent transplanting system
4.1
System working principle
(1) Working principle of Seedling recognition and transplanting By combining the designed transplanting device with the seedling identification scheme, the concept of an intelligent transplanting is developed, which is effective in the whole process of seedling transplanting. The working principle of the system is shown in Figure 7. According to the functional requirements, the system is mainly composed of seedling tray conveying mechanism, seedling picking mechanism, seedling recognition sensor and corresponding driven and control system. The frequency conversion motor is used to drive the picking mechanism and the planting mechanism to keep a certain rotational speed at 1:4 transmission ratio (that is, 60 to 120 seedlings/min) to match the motion. At the same time, the seedling tray conveyer mechanism is driven by step motor to supply the seedlings transversely and longitudinally, so as to realize the automatic picking, feeding and planting of the seedlings. In the whole process of seedling supply in the transverse and longitudinal direction, by detecting seedling identification, the seedling tray conveying mechanism stops waiting for the pot with seedling, and skip the pot without seedling on it and move to the next pot quickly. The mechanism motion matching and detecting and recognizing signal feedback in the whole process of intelligent transplanting are accomplished by PLC control system.
1
Move tray horizontally
Stepper motor
PlC controller
Speed regulating motor
Frequency converter
Pick and grab if seedling exist; Move to next grid to grab if no seedlings
Move tray longitudinally
2
Fig. 7
Conveying seedling
Receives seedling by planter
Release and drop
Carry seedling
3
Burrow and drop
4
Block diagram of working principle of intelligent transplanting system for automatic recognition of seedling vacancy
1, Computer logic control system; 2, Seedling tray conveying (automatic recognition of seedling); 3, Automatic seedling extraction mechanism; 4, Planting mechanism and seedlings dropping
(2) Structure composition of system for identifying Seedling The overall structure of automatic recognition system is shown in Figure 8. It is mainly composed of seedling tray conveying mechanism, five-bar and fixed axis gear train combined seedling picking mechanism and seedling recognition and control system. The seedling tray conveying mechanism is composed of a transverse and a longitudinal disk moving mechanism. The control system of seedling recognition is composed of the step motor, the detecting sensor, the claw position sensor, the initial position sensor, the left limit sensor and the program controller. Seedling picking mechanism performs a rotation movement in seedling extraction operation. When the position sensor is triggered, the program controller drives the stepper motor to rotate, and drives the conveying mechanism to move the seedling tray to the position to be taken. In the process of seedling supply, if the sensor detects that the pot has no seedlings, then the trigger program controller drives the step motor to rotate, and drives tray conveying mechanism to pass through the pot quickly to the next pot with seedling. Wait for the claw to grab, then complete picking, dropping, automatic recognition of the seedling. 11
10
9
1 Fig. 8
2
3
4
5
6
7
8
Automatic seedling recognition system
1, control cabinet 2,
body frame; 3, Stepping motor longitudinal move; 4, Seedling tray conveying mechanism; 5 6 10, Seedling detection
sensor; 7, Combined Seedling collecting Mechanism of Five-bar and fixed Axial Wheel Train; 8, Seedling claw position sensor; 9, Step motor for transverse moving; 11,Seeding initial position sensor
4.2
Design of seedling control system for recognition of seedling vacancy
According to the working characteristics of transplanting pepper seedling, the PLC programmable logic controller with strong adaptability and high reliability is used to develop the automatic control system for seedling vacancy identification. PLC control system consists of two parts: hardware and software. All kinds of logic control and time control are realized in PLC. 4.2.1 Hardware design The hardware of the control system mainly includes PlC controller, transverse moving step motor, longitudinal moving stepping motor、photoelectric sensor for seedling identification on the left side of transverse moving, photoelectric sensor for seedling identification on the right side of transverse moving, photoelectric sensor for seedling recognition on the left side of longitudinal moving, photoelectric sensor for identifying seedling on the right side of longitudinal moving, seeding initial position sensor, seedling claw position sensor, seedling tray conveying mechanism left limit sensor, etc. Its hardware composition is shown in Figure 9. Photoelectric sensor for identifying seedling deficiency on the left side of longitudinal feeding plate OPTEX(BGS-S08N)
Photoelectric Sensor for identifying seedling deficiency by right shift of Seedling Plate OPTEX(BGS-S08N) Photoelectric Sensor for identifying seedling deficiency by left shift of Seedling Plate OPTEX(BGS-S08N)
Seedling grab position sensor OMRON(E2E-Z-M10)
Left limit Sensor of Seedling Mechanism OMRON(E2E-Z-M10)
LED Integration Display DVP14SS211T(PLC)
Photoelectric sensor for identifying seedling deficiency on the right side of longitudinal feeding plate OPTEX(BGS-S08N)
Driver for step motor of moving horizontally Model SD-32208
Step motor for conveying seedling horizontally Model 110BYG350CHSAKSMA-0501
Driver for step motor of moving longitudinally Model SD-32208
Step motor for conveying seedling longitudinally Model 86BYG350CHSAKSHL-0301
Sensor for initializing seedling tray distance OMRON(E2E-Z-M10)
Fig. 9
Composition of Control system composition
In the control system, the input of PLC is connected with each photoelectric sensor and position sensor. And the output is directly connected with the driver of the transverse and longitudinal step motor to drive the operation of the stepping motor, so as to realize the seedling identification and orderly supply of the seedling. (1) Selection of PLC for programmable logic control PLC is responsible for collecting, analyzing and processing the output signal of the sensor in the system, and the output control signal is used to control the stepping motor to perform the corresponding action, which should be selected according to the number of I / O pins. The system requires at least 7 input points and at least 4 output points. According to the flexibility of programming and program debugging, expansion performance and performance-to-price ratio, the DVP14SS211T model of Delta Electronics is selected. It has 4 KB storage and 8 bit input and 6 pins output, two high speed pulse output terminals, built-in RS232 and RS485 communication port, can be connected with computer and LED
scrreen, and it eaasy to program m and debug. (2) Installaation of steppping motor an nd sensor Considerinng the need of differentt output torqque when trray conveyin ng mechanism m is used transversely t and lon ngitudinally, tthe 110BYG350CH-SAK KSMA-0501 ttype and 86B BYG350CH-S SAKSHL-03001 type threee-phase steppping mo otor produced by Hutchhison Compaany are usedd to drive the t transverse and longiitudinal shiftting mechannism resspectively. Thhere are two types t of senssor types usedd by the system. One is th he photoelecttric sensor ussed to detect and ideentify the seeedling. (4 BG GS-S08N back kground inhibbitory diffusee reflectance photosensorss produced by b OPTEX). The T oth her is a positiion sensor ussed to detect the position (the claw and the initial position p of thhe seedling mechanism). m T The req quired positioon sensor is innductor, whicch only deteccts metal objeects. Accordin ng to the sizee of the measu ured surface and thee required sennsing distance, we choosee the E2E-Z-M M10 type po osition sensor produced byy OMRON. According A to the fun nction of the stepping mootor and the detecting poosition requirred by the seensors, the sttep motor is installed on the transplanting deevice separateely, and the concrete distriibution is sho own in figure 10.
Fig g. 10
Installattion of sensorss and stepping motors
4.2 2.2 software ddesign (1) Control program floow Accordingg to the actionn requiremen nts of the inttelligent transsplanting con ntrol system, the flow chart of the onntrol sysstem for seeddling recognittion is proposed, and the coontrol system m program is composed, c ass shown in Fiigure 11.
Start system initialization Yes
No
Sensor 1 trigger? Yes Waiting for picking
Is the seedling supplying mechanism In the left limit? No
Yes Drive step motor to move longitudinally
Supplying Mechanism moves to the left limit
Tray moves longitudinalLy by 是 否 one grid
Drive step motor to move longitudinally Move the tray longitudinally to the picking position（tray distance switch signal detected）
Sensor 4 trigger? Yes
No
Exit sensor trigger？
Drive step motor to move horizontally
No
Sensor 3 trigger? Yes Waiting for picking
No
Exit sensor trigger？
Waiting for picking
No
Yes Exit sensor trigger？
No
Drive step motor to Yes move longitudinally
Yes
Tray moves to the left by one grid
Drive step motor to move horizontally Yes Tray moves to the right by one grid Yes
No Sensor 2 triggers?
Sensor 1 trigger?
Seven counts? Yes
No No Seven counts?
No
Sensor 2 triggers? Yes
Yes
No
Waiting for picking
Exit sensor trigger？
No
Yes No 15 counts?
Tray moves longitudinally by one grid
Drive step motor to move longitudinally
Yes End of the process
Fig. 11
Flow chart of control program for automatic identification of seedling
(2) Control system Circuit and Software Interface According to the program flow of the control system, the circuit diagram of the control system is designed, as shown in Figure 12, the I / O port of PLC is defined in Table 2. According to the circuit diagram of the control system, the man-machine interactive interface of the control system is designed (as shown in Figure 13), where two modes of operation (manual and automatic) can be switched, as well as pot distance, tray distance, number of rows of seedling tray, etc.
Fig g. 12
Circuit Diagram of Coontrol system for f automatic rrecognition an nd picking Seed dling
Table 2
I / O port definiition of PLC
Input
Output
X0
Leftmost lim mit (transverse) ssignal
Y0
Longitudinal step motor pulsse
X1
Seedling g tray limit signa nal
Y1
Longitudinal step motor pulsse
X2
Sk kip single
Y2
Transverse sstep motor pulsee
X3
Claw w exit signal
Y3
Transverse sstep motor pulsee
X4
X5
X6
X7
Signal 1 of riight shift detecttion of photoeelectric sensor Signal 2 of left l shift detectiion of photoeelectric sensor Signal 4 of ph hotoelectric sennsor for detection of th he right end of tthe tray Signal 3 of ph hotoelectric sennsor for detection of th he left end of thhe tray
a
Fig g.13
b
Control system interfaace
a Graphical G user innterface b Paraameter setting
5 Experimeent In order too test the quaality of the deesigned intellligent transpllanting system m, a comparaative experim ment was carrried outt in the laboraatory. The expperiment setu up was shownn in Figure 14 4, and the ind door illuminat ation was suffficient.
Fig g. 14 Indoor traansplanting exxperiment of seeedling identifiication
5.1 1 Condition n of experim ment In this expperiment, seeddlings at age of o 35d were uused for testin ng and the rate of emergennce of the seeedling was abbout 92% %. The speciific morpholoogical parameeters can be ffound in Secction 2.1. For statistics andd comparison n, the numberr of seeedlings in eacch tray was controlled c at 20 2 plants (thaat is, to contrrol the emerg gence rate to 85% for 128 8 pots). After the
tray was put into the conveying mechanism, the motor was started at 18.8Hz frequency to drive the continuous operation of the seedling picking mechanism and the rotary planting mechanism, corresponding to the transplanting speed of 90 plants /min. The whole experiment was divided into two groups. One group of control system did not turn on the function of automatic recognition and the other group turned on the function of automatic recognition. Each experiment was carried out until 10 trays were transplanted and the corresponding data were collected. 5.2
Test index
Important indicators to evaluate the performance of the transplanting system include the success rate of seedling acquisition and the rate of leakage. The definitions are as follows:
P=
T 100% T0
Eq. (1)
Where P is success rate of seedling acquisition; T is number of seedling acquisition; T0 is number of total seedlings.
D=
S 100% S0
Eq. (2)
Where D is leakage rate; S is number of dropping times; S0 is number of picking times. 5.3
Result of test
At the transplanting speed of 90 plants / min, the results of two groups with and without transplanting and identifying functions are shown in Table 3. In the Table 3, the number of seedling picking refers to the number of times that the seedling claw enters the pot during the process of seedling conveying. The number of successful seedlings picking is the number of seedlings that are successfully picked out from the pot with not damage (leaves and stems of seedlings have not been destroyed and more than 2 / 3 of the soil base is taken out). The number of successful seedlings dropping is the number of the seedling that are dropped into the planter with little damage (the stem and leaves of the seedlings are not broken and the soil base is not broken when contacting with the planter). Table 3 Results of transplanting experiments of two groups with and without the seedling recognition function Number
number of
The number
Number of
of
successful
of successful
seedlings
seedling
seedlings
seedlings
Transplanting
Successful
Working speed/plants· condition
min-1
Group 1： without the seedling
Tray No.
function
picking
picking
dropping
1
108
128
94
92
87.04
28.13
2
108
128
95
89
87.96
30.47
3
108
128
97
95
89.81
25.78
4
108
128
94
90
87.04
29.69
5
108
128
96
92
88.89
28.13
6
108
128
95
93
87.96
27.34
7
108
128
97
94
89.81
26.56
8
108
128
96
93
88.89
27.34
9
108
128
95
89
87.96
30.47
10
108
128
96
91
88.89
28.91
88.42
28.28
average value Group 2：with
recognition
rate/% picking
90
recognition
the seedling
Leakage rate of
90
1
108
111
98
89
90.74
19.82
2
108
107
93
90
86.11
15.89
3
108
105
94
90
87.04
14.29
function f
4
108
102
92
88
85.18 8
13.73
5
108
113
97
92
89.81
18.58
6
108
108
96
93
9 88.89
13.88
7
108
110
94
91
87.03 3
17.27
8
108
105
96
90
88.89 9
14.28
9
108
109
96
89
88.89 9
18.34
10
108
111
97
92
89.81
17.11
88.23 3
16.46
verage value av
5.4 4 Analysis aand discussioon The successs rate and leeakage rate in n two transpllanting statess were compaared and anallyzed when the t transplantting speeed was 90 / m min, as show wn in Figure 15.
a
Fig g. 15
b
Comparrison of transp planting resultss between two groups at diffeerent states wh hen transplantiing speed is 90 0 plants / min
a Picking P rate b Leakage rate
Accordingg to Fig. 15, under the co ondition of trransplanting speed s of 90 plants / min,, the automatic transplantting opeeration of the first group without the identificatioon function of the vacancy of seedlinggs was similaar to that of the intelligent transsplanting opeeration of the second grouup with the fu unction, and the success rrate of seedlings picking was w clo ose to that of the group 2. The averagee values of piicking rate were w 88.42% and a 88.23%, respectively. But the leakkage rate of the secoond group waas significanttly lower thann that of the first group, the t average rrate of the seecond group was w 28.28% and thee rate of the first f group waas 16.46%. T The main reasson is that the number of seedling rem moval is reducced, thaat is, the num mber of grabbiing soil base without seeddling is reduced, and the number n of tran ansplanting em mpty soil basse is red duced, whichh results in thhe decreased leakage ratee. Therefore, the designed d intelligent ttransplanting system can not onlly meet the reequirements of o high efficiiency transplaanting seedlin ng quality, bu ut also effectiively reduce the t phenomennon of seedling leakkage caused by b vacancy off seedlings; thhe leakage ratte is reduced by about 12% %. In the couurse of transsplanting exp periment, it is found thaat there are two kinds oof failure caases of seedlling ideentification: 1. There iss no seedling in the pot bu ut before the seedling claw w moves into the hole, thee seedling traay conveyer does d nott remove trayy in correct position, p whicch leads to thhe failure grab bbing. The main m reason fo for this situatiion is the sennsor misjudgment caaused by the interference i of o neighborinng seedlings. 2. There arre seedlings in i the pot but before the cllaws enter thee hole, the seedling tray coonveyer mov ves away the pot, p wh hich results inn the miss of grabbing g thiss seedling. Thhe main reaso on for this pheenomenon is that the seedling is entanggled witth neighborinng seedling or o the parts of o the machinne, which caauses the sen nsor to be unnable to identtify the seedlling
because of the sensitivity limitation of the sensors. Table 3 shows that the number of seedling picking is less than the total number of seedlings. Although this phenomenon had no effect on the leakage rate of the system, it caused a waste problem and affected the economy of seedlings transplanting. In view of the above problems, we can improve the design by adjusting the sensitivity of the photoelectric sensor and the tilt angle of installation, controlling suitable moisture content of the potted seedling (water content directly affects the upright degree of potted seedling during operation), removing the parts (pressure disc rod, seedling leaf stem) that may be wound with the leaf stem of the seedling on the condition of not affecting the working function of the mechanism, etc. Measures can be taken to improve the accuracy of automatic seedling identification, and reduce the rate of seedling leakage and further control the level of seedling waste.
6
Conclusions
(1) An intelligent transplanting system was proposed on the basis of the transplanting device for potted pepper seedlings. The system is composed of seedling picking mechanism with the combination five-bar mechanism and fixed axle train, seedling tray conveying mechanism, rotary planting mechanism, photoelectric sensor for seedling detection, position sensor, stepper motor and PLC control system. By the use of computer logic programming, the motor drives each mechanism in a coordinated and orderly manner, and it can detect whether the soil base with or without a seedling on it, and realize the automatic recognition of the seedling. (2) A comparative experiment on transplanting for pepper seedlings in laboratory showed that: when the transplanting speed is 90 plants/min, the intelligent system can effectively identify the condition of the soil base, so as to avoid picking the defect soil base and reduced leakage rate caused by the vacancy of seedlings; The success rate of seedling picking was 88.23% and the leakage rate was 16.46%. Compared with the transplanting model without seedling identification, the average rate of leakage was reduced about 12%.
Acknowledgements The work was sponsored by the National Key Research and Development Program of China Sub-project (No.2016YFD0700103 & No.2017YFD0700800), and the Innovation Scientists and Technicians Troop Construction Projects of Henan Province (No.184200510017).
[References] [1]
[2]
[3]
[4]
[5]
[6]
Wu J M., Zhang X C., Jin X., et al.: Design and experiment on transplanter pot seedling disk conveying and positioning control system. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE). 31(1), 47-52 (2015). (in Chinese with English abstract) Jin X., Du X W., Ji J T., He Z T., et al.: Automatic detection and controlling system of pot seedlings for transplanters. J. International Agricultural Engineering Journal. 24(2), 143-151 (2015). Yang Z Y., Zhang W Q., Li W., et al.: Monocular vision-based method for direction adjustment of transplanting potted-seedling leaves. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE). 30(14), 26-33(2014). (in Chinese with English abstract) Tai Y W.，Ling P P.，Ting K C.: Machine vision assisted robotic seedling transplanting. J．Transactions of the ASAE. 37(2), 661-667 (1994)． K H Ryu.，G Kim.，J S Han.: Development of a robotic transplanter for bedding plants. J．Journal of Agricultural Engineering Research. 78(2)，141-146 (2001)． Humphries S.，Simonton W.: Identification of plant parts using color and geometric image data. J．Transactions of the ASAE. 36(5), 1493-1500 (1993)．
[7] [8] [9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
Mizuochi Y，Dohi M.: Machine vision for transplanter of vegetables. C．ASAE Annual Meeting，2003． Albertus J V V.: Multiple Transplanting Apparatus[P]．US，No．20040020110A1，(2004)． HU F.，YIN W Q.，CHEN C R.，XU B.: Recognition and localization of plug seedling based on machine vision. J. Journal of Northwest A＆F University(Nat．Sci．Ed． 41(5), 183-188 (2013). Hou S.: Research on Methodology to identify Seedling Health Condition of Plug Seedling Transplanting Robot. Master Thesis, Jiangsu University, (2015). Yang Z Y., Zhang W Q., Li W., et al.: Information acquisition method of potted-seedling transplanting fitness using monocular vision. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 30(3), 112－119 (2014). (in Chinese with English abstract) He L Y., Yang T W., Wu C Y.: Optimization of Replugging Tour Planning Based on Greedy Genetic Algorithm. J. Transactions of the Chinese Society for Agricultural Machinery. 48(5), 36-43 (2017). (in Chinese with English abstract) Liu J D., Cao W B., Xu H Z., et al.: Adaptive fuzzy-PID control of accurate orientation for auto-detect seedling supply device. J. Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE), 33(9), 37－44 (2017). (in Chinese with English abstract) Jin X., Du X W., Ji J T., et al.: Physical characteristics of plug seedling transplanted by the return-blank type transplanter. J. International Agricultural Engineering Journal, 24(4), 1-10 (2015). Wang Y Y., Yu H Y., Liu Y Y.: Mechanical transplanting plug tray localization method based on binocular stereo vision. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 32(5), 43-49 (2016). (in Chinese with English abstract) Ni Y L., Jin C Q., Liu J.: Design and experiment of system for picking up and delivering seedlings in automatic transplanter. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE). 31(23), 10-19 (2015). (in Chinese with English abstract) Sun G X.: Automatic plug seedlings Transplanting Robot Based On Machine Vision. Master Thesis, Nanjing Agricultural University. (2009) Tan M H.，Cao Q X.，Qiu Q.，et al.: Real-time Localization Research of Transplanting Based on Machine Vision. J. Transactions of the Chinese Society for Agricultural Machinery. 47(2), 59-63 (2016). Jin X., Du X W., Ji J T., et al.: Mechanical property experiment of plug seeding with pots gripping-picking. J. International Agricultural Engineering Journal, 24(4), 24-33 (2015). Feng Q C., Wang X., Jiang K., et al.: Design and test of key parts on automatic transplanter for flower seedling. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 29(6), 21-27 (2013). (in Chinese with English abstract) Yang Z Y., Zhang W Q., Li W., et al.: Monocular vision-based method for direction adjustment of transplanting potted-seedling leaves. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE). 30(14), 26-33 (2014). (in Chinese with English abstract). Yang Y., Gao Q X., Sheng G D.，et al.: Plug Tray Localization and Detection System Based on Machine Vision. J. Transactions of the Chinese Society for Agricultural Machinery. 44(6), 232-235 (2013). (in Chinese with English abstract) Lu C H.，Chen X G., Wu W F.，et al.: Automatic Measuring System of Seedling Perpendicularity Based on Binocular Stereo Vision. J. Transactions of the CASE. 33(1), 60-62 (2002). WANG Y Y.，YU Y.，LIU Y Y.: Recognition of seedlings in mechanical transplanting processing by Morphological image detection. J. Optics and Precision Engineering. 24(3), 651-658 (2016). FENG L., CHAI R Y.，SUN G M.，et al.: Identification and Classification of Rice Leaf Blast Based on Multi--Spectral Imaging Sensor. J. Spectroscopy and Spectral Analysis. 29(10), 2730-2733 (2009). Deng W., Huang Y B., Zhao C J., Wang X.: Identification of seedling cabbages and weeds using hyperspectral
imaging. J. Int J Agric & Biol Eng. 8(5), 65-72（2015）. [27] Narendra V G., Hareesh K S.: Study and comparison of various image edge detection techniques used in quality
[28]
[29]
[30]
[31]
inspection and evaluation of agricultural and food products by computer vision. J. Int J Agric & Biol Eng. 4(2), 83-90 (2011). Wu Q., Wang C., Fang J J., Ji J W.: Field monitoring of wheat seedling stage with hyperspectral imaging. J. Int J Agric & Biol Eng. 9(5), 143-148 (2016). Wang Q H., Tang Y H., Xiao Z.: Grape size detection and online gradation based on machine vision. J. Int J Agric & Biol Eng. 10(1), 226-233 (2017). Li B R., Long Y., Song H B.: Detection of green apples in natural scenes based on saliency theory and Gaussian curve fitting. J. Int J Agric & Biol Eng. 11(1), 192-198 (2018). Jin X., P J., Ji J T., et al.: Experiment and simulation analysis on high-speed up-film transplanting mechanism. J. International Agricultural Engineering Journal, 26(3), 105-112 (2017).
Jin Xin, PhD, Associate Professor, College of Agricultural Equipment Engineering, Henan University of Science and Technology. Dr. Jin focuses on application of computer programming and sensor Technology in Agriculture.
Zhao Kaixuan, PhD, Lecturer, College of Agricultural Equipment Engineering, Henan University of Science and Technology. Dr. Zhao focuses on artificial intelligence and image processing algorithm. Ji Jiangtao, PhD, Professor, College of Agricultural Equipment Engineering, Henan University of Science and Technology. Prof. Ji focuses on the study of intelligent and information technology of agricultural machinery equipment. Du Xinwu, PhD, Associate Professor, College of Agricultural Equipment Engineering, Henan University of Science and Technology. Dr. Du focuses on design of agricultural machinery and analysis of field experiment Ma Hao, PhD, Lecturer, College of Agricultural Equipment Engineering, Henan University of Science and Technology. Dr. Ma focuses on application of big data analysis and information processing technology in agriculture. Qiu Zhaomei, PhD, Associate Professor, College of Agricultural Equipment Engineering, Henan University of Science and Technology. She focuses on the study of computer
programming and control system.
Jin Xin
Zhao Kaixuan
Ji Jiangtao o
Du Xin nwu
Ma Hao
Qiu Zhaomei Z
Research Highlights: 1. Propose the idea of intelligent transplanting in field planting of agricultural production. 2. On the basis of computer logic programming, combined crop seeding, photoelectric sensor and mechanical system, we developed an intelligent control system to realize the active recognition and effective collection of seedling in the transplanting process. 3. The average rate of missed planting of the machine with intelligent control system was reduced about 12%, which has the potential to solve the miss planting problem caused by vacancy of seedlings.
PII: DOI: Reference:
S0167-739X(18)30634-4 https://doi.org/10.1016/j.future.2018.05.034 FUTURE 4209
To appear in:
Future Generation Computer Systems
Received date : 22 March 2018 Revised date : 23 April 2018 Accepted date : 15 May 2018 Please cite this article as: J. Xin, Z. Kaixuan, J. Jiangtao, D. Xinwu, M. Hao, Q. Zhaomei, Design and implementation of Intelligent transplanting system based on photoelectric sensor and PLC, Future Generation Computer Systems (2018), https://doi.org/10.1016/j.future.2018.05.034 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Design and implementation of Intelligent transplanting system based on photoelectric sensor and PLC Jin Xin1,2 Zhao Kaixuan1 Ji Jiangtao1,2* Du Xinwu1 Ma Hao1
Qiu Zhaomei1
1. College of Agricultural Equipment Engineering, Henan University of Science and Technology, Luoyang 471003, China; 2. Collaborative Innovation Center of Machinery Equipment Advanced Manufacturing of Henan Province, Luoyang 471003, China. * Corresponding author: Ji Jiangtao, Professor, dedicated to the research on new technology and theory of agricultural machinery and intelligent agricultural equipment. Address: 48 Xiyuan Road, Luoyang 471000, Henan Province, China. Tel:+86-379-64877837, Email: [email protected] Abstract: Industrial seedling rearing is an important part of seedling transplanting technique. At present, there exists the phenomenon that soil base on the tray may have empty seedling on it, which will lead to a large number of leakage and reduce the yield in the course of automatic transplanting in the field. To solve this problem, automatic seedling identification method is studied to skip picking empty soil bases. An intelligent transplanting system are designed with the use of the picking mechanism five-bar and fixed-axis gear train, the seedling tray conveying mechanism (with transverse and longitudinal seedlings feeding function), the eccentric disc parallel four-bar duck mouthpiece planting mechanism, the electric sensor for seedling detection and identification of the seedling tray, the position sensor, stepper motor and PLC control system. By use of computer logic programming and control, it can be used to identify effectively whether the soil base is empty or not, and to control the automatic transmission of seedling tray. The tray is moved to quickly skip the action path of the manipulator to avoid the picking of empty soil base. The system was tested by using pepper seedlings on indoor condition. The results showed that when the transplanting frequency was 90 / min, the rate of success seedling picking was 88.23% and the leakage rate was 16.46%, which could meet the requirement of pepper transplanting. Compared with the transplanting model without seedling identification, the average rate of leakage was reduced about 12%. The research provides a useful reference for the development of intelligent agricultural transplanting technology and equipment. Keywords:
1
Intelligent agricultural; Transplanting system;
Automatic recognition;
PLC control;
Logic programming
Introduction
The vacancy of seedling in factory seedling cultivation is an important factor leading to the leakage of subsequent mechanical transplanting. In the whole automatic transplanting process, the number of seedlings planted in the unit area will be reduced if the vacancy of seedling cannot be effectively identified and avoided which will reduce the production. At present, the main way to deal with this problem is to fill the seedlings after planting. The whole process is characterized by high labor intensity, low efficiency and high cost [1-3]. Therefore, it is urgent to improve the intelligent level of mechanical transplanting technology. The advanced sensor and computer logic control technology are applied to the transplanting equipment to realize the independent identification and selective operation of the machine. Many useful researches have been done in the field of seedling identification and intelligent transplanting. For example, Tai et al. [4] used machine vision for automatic seedling raising in greenhouse, a recognition system is developed, which can automatically identify the empty pot on the seedling tray. Ryu et al. [5] designed an automatic transplanter system including machine vision, which can realize the functions of automatic conveying, grabbing and planting seedling. Humphries S et al. [6] used the image tracking algorithm to divide the seedling plants into different parts with similar geometric characteristics, and realize the classification and recognition of the leaves, stems and main *
Corresponding author: Ji Jiangtao, Professor, dedicated to the research on new technology and theory of agricultural machinery and intelligent agricultural equipment. Address: 48 Xiyuan Road, Luoyang 471000, Henan Province, China. Tel:+86-379-64877837, Email: [email protected]
stems of the seedlings. Mizuochi et al. [7] used machine vision technology to develop the information recognition system of the seedling, according to the shape characteristics of the cotyledon leaf circumference of the seedling, to identify the healthy seedling, and to eliminate the diseased and abnormal seedlings. Albertus et al. [8] designed the multi-disc automatic conveying device, and based on the vision recognition system, realized the automatic seeding supplying and seeding feeding; Hu Fei et al. [9] designed a machine vision system for transplanting seedlings, which was used to measure the leaf area of seedlings in each pot in real time, and to determine the suitable grasping position for transplanting seedlings. Hou et al. [10] studied the identification technology of healthy seedling, and established a comprehensive evaluation method based on multi image feature of healthy seedling, which can realize the identification of robust seedling and the location of rootstock. Yang et al. [11] based on monocular vision technology, designed an algorithm to obtain the information of transplanting suitability of potted seedlings, and used the Harris corner detection algorithm to obtain the information of main stem of potted seedlings. He et al. [12] studied the method of optimizing the path of the seedling supplying, and used machine vision to obtain the information of the seedlings, to control the executer to grab the healthy seedlings to complete the transplanting. In order to realize the precise positioning control of transplanting and seeding device, the adaptive fuzzy PID controller and its control rules were designed by Liu et al. [13] However, most of the above studies are focused on the process of seedling raising in factories. The health status identification, pot location and culling control technology of seedling leakage are seldom reported, and few reports have been made on the control technology of the transplanter to identify and evade the seedling vacancy in the pot tray. In this paper, in order to solve the problem of self-identification and avoiding the vacancy of seedling in transplanter, this paper analyzes and studies the seedling identification and positioning control of automatic transplanting machine, and intelligent transplanting system is developed with mechanical design, the sensor, computer programming and PLC control, etc. The system recognizes the vacancy of seedling by photoelectric sensor, localizes the position of the pot with seedlings vacancy by combining the position sensor, and transmits the position information to the PLC controller to guide the seedling feeding mechanism to evade that pot, to achieve the goal of precise seedling extraction by the seedling manipulator. Using photoelectric sensor, position sensor and computer logic programming to replace machine vision to obtain the information of seedling vacancy, which avoids the interference of the machine vision caused by the complex working environment of transplanting, and can provide a reference for the intelligent design of transplanting equipment.
2
Structure and working principle of transplanting device
2.1
Transplanting object
The transplanting object is the pepper seedlings, and the variety is Hejiao 12, which originated from Mexico. The seedling plate is a conical tray of 128 pots. The depth of the pot is 41.5mm. The spacing between pots is 36.5mm. The of size of the pot top is 30.5mm × 30.5mm. The size of the pot bottom is 15mm × 15mm, and the volume of the cavern is about 21ml. The seedling substrate was mainly peat and vermiculite with a volume ratio of 3: 1; The way of seedling raising is to adopt factory cultivation, that is, precision seeding equipment is used to complete filling the pot, seeding, covering and compacting in one operation[14-18], then replenishing water regularly, controlling temperature and light, etc. The seedlings at age of 35 days are suitable for transplanting. At this time the seedlings grow 5~6 leaf and root has matrix winding package, as shown in Figure 1a; The industrial seedling breeding is based on the idea of precision seeding, that is, one seed for one pot. Although the environment, water, fertilizer and other crop growth conditions of seedling raising process are standardized, but because of the difference of seed quality and artificial control growth conditions, there is no guarantee that every seed can emerge. At present, the seedling emergence rate of Chinese factory is over 90%, as shown in Figure 1b. The average diameter of the seedling stem was 3.56mm (about 10mm from the top surface of the seedling pot), the height of the seedling stem was 40mm ~ 51mm, and the moisture content was about 52%.
a
Fig g.1
b
Seedlings at age of 35 daays
a Single S potted seeedling.
d, Diaameter of stem; h, Height of steem; l, Seedling leaf broadening
b Seedling tray with vacan ncy
2.2 2 Structuree of transplan nting device The transpplanting devicce is mainly composed c off seedling tray y conveying mechanism, m aautomatic seeedling collectting meechanism andd rotary plannting mechaanism, as shoown in Figu ure 2. The design d of auutomatic seedling collectting meechanism refeers to the com mbined plantin ng-taking meechanism of five-bar f and fixed f axis geaar train [19]. Itt mainly conssists of gear transmiission box (a pair of straig ght cylindric al gears of th he same size meshing), ddouble crank, connecting rod, r seeedling arm, seeedling claw (a pair of seeedling needlee, push rod and a ring) and other parts ((as shown in Figure 2b). The T two o crank are ddriven by gearr to rotate in the opposite direction, and the power is i transferredd from the con nnecting rod and according to the predeterm thee seedling arm m to the seedlling claw, so that the seeddling claw is reciprocated r mined track, and thee seedlings arre taken and dropped d in th he fixed positiion. The desiign requiremeent of seedlinng tray conveying mechannism is that t each pott on the tray can be moved intermitteently to the position p to bee taken accorrding to a cerrtain order. This T pro ocess avoids interference between seedling tray annd seedling claw, c and sho ould be conveenient to recover empty tray t after taking seeedling. The design d adopts two parallell synchronouss chain drivees to realize tthe longitudin nal tray feediing. Th he guide rail llead screw mechanism waas used to carr rry the seedlin ng tray transv versely. Theyy are all driveen by step mootor, and d the distance to move is equal to the distance bettween pots, which w is 36.5 mm (see Figgure 2a). Thee rotary plantting meechanism adoopts duckbillled planting mechanism with eccenttric disc and d parallel fouur-bar [20], which w is maiinly com mposed of foour duckbill planters, cam m sliders, ecccentric diskss, left and rig ght plates annd eccentric disc slides. The T meechanism cann realize rotaryy planting an nd keep the duuckbill planteer always verttical downwaard.
Fig g.2
Plant tran nsplant device for pepper seeedling
a Tray T conveying mechanism b Combined C Seed dling collecting Mechanism of Five-bar and fiixed Axial Wheeel Train. 1, electric e control cabinet; 2, boddy frame; 3, ducckbilled plantingg mechanism with w eccentric diisc and parallel four-bar; 4, converter motor 5, seedling s claw; 66, Seedling arm m; 7, double cran nk; 8, connectinng rod; 9, driven gear; 10, Two o parallel synchhronous driven chain; 11, Seedling tray; 12, Guide rail screw s mechanissm; 13, transverrse moving step p motor
2.3 3 Working principle of transplantin ng device The autom matic control technology is used to connnect the mottion of seedliing conveyinng, picking an nd planting. The T wo orking principple block diaggram of the trransplanting ddevice for pep pper seedling gs is shown ass shown in Fiigure 3. a Move the tray y
Stop at the ppicking
horizontally
positionn
Finish F serving seedlings in row horizontally h
Yes Move e the tray once to the picking
No
position
seedlingg tray conveying mecchanism
b
Autom matic seedling pickingg mechanism Picking the t seedling
c
Pla anting mechanism
Burrow and drrop
by y claw
Conveeying the
Carry seedlin ng
seeedling
Receive seedling by Releasee and drop
the planter
Fig.3 Working principle of transplanting device a Seedling tray conveying step b Automatic seedling taking c Seedling planting step
The seedling taking mechanism and the rotary planting mechanism always do reciprocating motion according to a certain path, and the power is provided by frequency conversion motor and driven chain. The transmission ratio of two mechanisms is set to 1: 4, that is, each time the seedling claw reaches the lowest point (the position of sapling), the corresponding duckbill transplanter rises to the highest point (seedling position) and matches it. The transverse and longitudinal feeding of the seedling tray by conveyer mechanism is powered by stepping motor. When the seedling is taken out completely, the step motor drives the seedling tray to move to the next pot transversely. The claw moves downward with the seedling, and when moves to the dropping position, the duckbilled planter rises to the position where the seedling is received, and the seedling claw releases the seedling and place it into the planter. The planter then moved down until the lowest point was completely opened, and implanted the seedling into the field. At the same time, the planter continued to grab the next seedling which had been moved to the position for picking up [21-24]. In this way, when the seedling grabs the last seedling of a row, the step motor drives the tray to move to the next pot longitudinally, so that the next row of pots on the edge of the tray is moved to the position to be picked up, and then continue to complete the feeding, taking, and planting. The whole process moves in cycles, until the whole tray of potted seedling were transplanted.
3
Design of vacancy recognition scheme for seedling
3.1 Requirements for detection and identification of seedlings The intelligent transplanting of pepper seedling requires that it should be able to identify whether there is a vacancy of seedlings or not in the seedling collection stage, and to grasp the seedlings, and move quickly to the next seedling position if there is no seedling. The following conditions should be satisfied in order to realize the rapid detection and identification of the seedlings: 1. After the seedlings were raised in the pot, the top branches and leaves of the seedlings were mostly cross-occluded, which make it difficult to distinguish the branches and the leaves of the seedlings [25,26]. This problem should be solved or avoided when the sensor detects whether there is a seedling or not. 2. The interference of adjacent seedlings should be overcome. 3. Not subject to restrictions on testing materials. 4. Sensor with high accuracy and sensitivity is adjustable. 5. Sensor detection and identification signal feedback ha high speed and is easy to use. 6. Sensor structure size is small; light is weight, and easy to install and adjust position. 3.2
Selection and verification of the sensor for seedling detection
(1) Sensor selection In order to meet the above conditions, on the basis of analyzing the researches on the detection and identification of nonmetallic bodies, combining with the structure and operation characteristics of the transplanting device, BGS-S08N background inhibitory diffuse reflectance photoelectric sensor is used to detect and identify the pepper seedlings [27-30]. The specific parameters are shown in Table 1. The photoelectric sensor integrates a transmitter and two receivers, with simple structure, small size and easy installation; At the same time, it has the function of background suppression, which can effectively avoid the interference of strong background light beyond the detection distance, which is more reliable than the conventional reflective sensor detection. The sensitivity can be adjusted, which make it easy to adjust the detection distance.
Table 1
Model
Dimensioons
Technical pparameters of th he photoelectricc sensor
deteection
Dettection
mode
disstance //mm
onse respo
voltag ge
current
Control
levels of
time /ms
/VDC C
/mA
output
protection
<30
NPN
IP67
Backg ground BGS-S08N N
supprression difffuse
100~80
<0.5 0
0 10~30 (±10% %)
refleection
(2) Feasibiility Test of Seedling S vacancy in pots In order too verify the feasibility off BGS-S08N background inhibitory diffuse reflecttance photoelectric sensorr to dettect and identtify the planttlets, the senssitivity of the detection an nd identificatiion of non-me metallic materiials was studied. Th he response tim me and the acccuracy of seedling stem rrecognition under natural light l were tessted. Test methood: the seedliing tray is fix xed, the photooelectric senssor is parallel to the tray suurface. The sensor s is movving at a certain speeed to detect thhe seedling sttem. The test coonditions weere as followss: 3 tray of ppepper seedliings were sellected and soome seedlings were removed ran ndomly to maake 40 vacanncy seedlingss; It was testted under sim mulated naturral light in thhe laboratory y. The transveerse disstance betweeen photoelecttric sensor an nd seedling sttem is about 35mm, and the vertical diistance is abo out 15mm to the surrface of seedlling bowl. Thhe sensors weere moved traansversely baack and forth at a speed off above 4 potts / s. After each e row w, the photoeelectric sensoor moves forw ward to the nnext row and continues to detect. The rresponse of the t sensor to the tray is shown inn Figure 4. The resultss showed thatt 320 pots in 40 rows, andd 40 vacancy seedlings weere detected inn the experim ment. The triggger reaaction speed of the sensoor is relatively fast (by thhe output sign nal indicator light). All oof pot with seedlings s cann be ideentified, and tthere are 2 vaacancy seedlin ngs failed to detect, and th he accuracy of o recognitionn of vacancy seedling s is 955%. The main rreasons for missing m detecttion are the innterference of adjacent seeedlings, suchh as the tilting g of the right and lefft seedlings, aand the lodgiing of the seedlings in thhe back row, which result in the seedliing leaves bllocking the liight beaams of the sensors and cauusing misjudg gment.
Fig g.4
Response Test of BGS-S S08N photoelecctric Sensor
Accordingg to the expeerimental results, it can bbe concluded d that it is feasible f and effective to use BGS-S008N pho otoelectric seensor to idenntify the seeedling stem. Then in thee transplantin ng device, foour BGS-S08 8N photoelecctric sen nsors are arraanged separaately (as show wn in Figuree 6) for autom matic identiffication of thhe whole seedling conveyying pro ocess. The seeedling will be b taken out when the pott with seedlin ng. Otherwise, the pot wiill be abando oned. The dessign ideea of movingg the tray to the next seeedling pot whhen vacancy is detected should have the same feasibility, andd its imp plementationn effect needs to be verified by experim ment. 3.3 3 Identificaation schemee of seedling vacancy durring seedling g tray convey ying The installlation height of the BGS--S08N photoeelectric senso or was adjustted to detect the seedling stem below the misjudgmentt caused by the leaaf (about 10m mm / 15mm m from the top surface oof the pot), which could d avoid the m
interleaving of the leaves. The T sensitivity y of the sensoor can be adjjusted to overrcome the intterference off the seedlings in thee back row. Figure 5 is a schematic diag gram of sensoor detection and a identificaation. It can bbe seen from the diagram that t no matter wherre the seedlinng position is taken, the beam emitted by the sen nsor must alw ways be perp pendicular to the seeedling stem, that is, paralllel to the su urface of the pot. At the same time, th he sensor shhould have a certain distaance relative to the sseedling stem m, to prevent the t interferennce of the seeedling in the back row andd the interferrence of the trray. In the process oof the seedling tray convey yor moving thhe seedling to o the seedling g position, thhe seedling steem goes throuugh thee beam of thee sensor, and the sensor is triggered to pprove that th here is a seedlling. If the seensor is not trriggered, therre is no seedling.
Fig g 5.
Schematiic Diagram of Detection D and recognition off Seedling
1, BGS-S08N B phootoelectric sensoor;
2, Seedling g tray;
3, Lonngitudinal shift driven chain; 4, Long pin shhaft
Because thhe mechanism m of seedling g supply neeeds to supply the seedling gs both transvversely and longitudinally l y, it neeeds the coopeeration of fouur photoelectrric sensors too complete th he recognition n of seedling vacancy [2, 13, 31]. The speccific arrrangement off the sensors inn the automatic transplantting device is shown in Fig gure 6.
Fig g. 6 Installation n of photoelecttric sensors 1, Stepping S motor for longitudinaal move; 2 (6 10 0 12), BGS-S088N photoelectriic sensor; 3, Meechanism for loongitudinal mov ve; 4, Mechanissm for transverse movve; 5, Seedling tray; 7, Guide rail r screw mechhanism; 8, Stepp ping motor for transverse movve; 9, Seedling claw c position sen nsor; 11, Seedlinng claw
It can be seen from Figure 6 that the implementation of seedling vacancy identification in the tray conveying process mainly depends on the transverse and longitudinal movement of step motor and four BGS-S08N photoelectric sensors. In the picture, the seedling tray is located at the leftmost end of the mechanism. The seedling claw is pointing to the pot No. 1. Sensor 12 and sensor 10 are on the left and right sides of the seedling claw, and placed in the right side of the longitudinal centerline of pot 2 and pot 1 which are fixed to the frame. The position of sensor 6 and sensor 2 are placed under the lateral centerline of pot 1 and pot 8 (fixed with the machine frame and moving with transverse movement). The realization process of seedling vacancy recognition is as follows: 1. The tray is fed longitudinally and the pot 1 is moved to the seedling position and the longitudinal motion motor stops. If the sensor 6 detects that there is a seedling in the pot cell 1 in the process of moving forward, then it waits for the seedling to be picked up, and enter the step 2 the once the seedling is taken out; If sensor 6 does not detect a seedling in pot cell 1, then entered step 2 immediately. 2. Move the tray to the next pot transversely, and move the pot 2 to the position to be taken. Wait for the seedling claw to take the seedling, and move the tray to the next right pot after taking out. In the process of right shift, if the sensor 12 does not detect the seedlings in the next pot, then immediately move the next pot to the position to be taken. 3. Repeat step 2 until the seedling in the pot 8 in the tray is taken out. 4. Move the tray to the next pot longitudinally, and move the pot 9 to the position to be taken, then stop the longitudinal motion motor. If the sensor 2 detects the presence of seedlings in pot 9 as it moves forward, it waits for the claw to taking the seedlings, and enter step 5 after the seedlings are taken out. If sensor 2 does not detect seedlings in cell 9, then enter step 5 immediately. 5. Move the tray to the next left pot, and move pot 10 to the position for taking out. Wait for the claw to pick up seedling, and move the tray to the next left pot. During the left shift, if the sensor 10 does not detect a seedling in the pot, then move to next left pot immediately, and move the next pot to the position to be taken. 6. Repeat step 5 until the seedling in pot 16 is taken out. 7. Move forward longitudinally, and move pot 17 to the position to be take, and then repeat step 2 until the whole seedlings are taken out.
4
Design of intelligent transplanting system
4.1
System working principle
(1) Working principle of Seedling recognition and transplanting By combining the designed transplanting device with the seedling identification scheme, the concept of an intelligent transplanting is developed, which is effective in the whole process of seedling transplanting. The working principle of the system is shown in Figure 7. According to the functional requirements, the system is mainly composed of seedling tray conveying mechanism, seedling picking mechanism, seedling recognition sensor and corresponding driven and control system. The frequency conversion motor is used to drive the picking mechanism and the planting mechanism to keep a certain rotational speed at 1:4 transmission ratio (that is, 60 to 120 seedlings/min) to match the motion. At the same time, the seedling tray conveyer mechanism is driven by step motor to supply the seedlings transversely and longitudinally, so as to realize the automatic picking, feeding and planting of the seedlings. In the whole process of seedling supply in the transverse and longitudinal direction, by detecting seedling identification, the seedling tray conveying mechanism stops waiting for the pot with seedling, and skip the pot without seedling on it and move to the next pot quickly. The mechanism motion matching and detecting and recognizing signal feedback in the whole process of intelligent transplanting are accomplished by PLC control system.
1
Move tray horizontally
Stepper motor
PlC controller
Speed regulating motor
Frequency converter
Pick and grab if seedling exist; Move to next grid to grab if no seedlings
Move tray longitudinally
2
Fig. 7
Conveying seedling
Receives seedling by planter
Release and drop
Carry seedling
3
Burrow and drop
4
Block diagram of working principle of intelligent transplanting system for automatic recognition of seedling vacancy
1, Computer logic control system; 2, Seedling tray conveying (automatic recognition of seedling); 3, Automatic seedling extraction mechanism; 4, Planting mechanism and seedlings dropping
(2) Structure composition of system for identifying Seedling The overall structure of automatic recognition system is shown in Figure 8. It is mainly composed of seedling tray conveying mechanism, five-bar and fixed axis gear train combined seedling picking mechanism and seedling recognition and control system. The seedling tray conveying mechanism is composed of a transverse and a longitudinal disk moving mechanism. The control system of seedling recognition is composed of the step motor, the detecting sensor, the claw position sensor, the initial position sensor, the left limit sensor and the program controller. Seedling picking mechanism performs a rotation movement in seedling extraction operation. When the position sensor is triggered, the program controller drives the stepper motor to rotate, and drives the conveying mechanism to move the seedling tray to the position to be taken. In the process of seedling supply, if the sensor detects that the pot has no seedlings, then the trigger program controller drives the step motor to rotate, and drives tray conveying mechanism to pass through the pot quickly to the next pot with seedling. Wait for the claw to grab, then complete picking, dropping, automatic recognition of the seedling. 11
10
9
1 Fig. 8
2
3
4
5
6
7
8
Automatic seedling recognition system
1, control cabinet 2,
body frame; 3, Stepping motor longitudinal move; 4, Seedling tray conveying mechanism; 5 6 10, Seedling detection
sensor; 7, Combined Seedling collecting Mechanism of Five-bar and fixed Axial Wheel Train; 8, Seedling claw position sensor; 9, Step motor for transverse moving; 11,Seeding initial position sensor
4.2
Design of seedling control system for recognition of seedling vacancy
According to the working characteristics of transplanting pepper seedling, the PLC programmable logic controller with strong adaptability and high reliability is used to develop the automatic control system for seedling vacancy identification. PLC control system consists of two parts: hardware and software. All kinds of logic control and time control are realized in PLC. 4.2.1 Hardware design The hardware of the control system mainly includes PlC controller, transverse moving step motor, longitudinal moving stepping motor、photoelectric sensor for seedling identification on the left side of transverse moving, photoelectric sensor for seedling identification on the right side of transverse moving, photoelectric sensor for seedling recognition on the left side of longitudinal moving, photoelectric sensor for identifying seedling on the right side of longitudinal moving, seeding initial position sensor, seedling claw position sensor, seedling tray conveying mechanism left limit sensor, etc. Its hardware composition is shown in Figure 9. Photoelectric sensor for identifying seedling deficiency on the left side of longitudinal feeding plate OPTEX(BGS-S08N)
Photoelectric Sensor for identifying seedling deficiency by right shift of Seedling Plate OPTEX(BGS-S08N) Photoelectric Sensor for identifying seedling deficiency by left shift of Seedling Plate OPTEX(BGS-S08N)
Seedling grab position sensor OMRON(E2E-Z-M10)
Left limit Sensor of Seedling Mechanism OMRON(E2E-Z-M10)
LED Integration Display DVP14SS211T(PLC)
Photoelectric sensor for identifying seedling deficiency on the right side of longitudinal feeding plate OPTEX(BGS-S08N)
Driver for step motor of moving horizontally Model SD-32208
Step motor for conveying seedling horizontally Model 110BYG350CHSAKSMA-0501
Driver for step motor of moving longitudinally Model SD-32208
Step motor for conveying seedling longitudinally Model 86BYG350CHSAKSHL-0301
Sensor for initializing seedling tray distance OMRON(E2E-Z-M10)
Fig. 9
Composition of Control system composition
In the control system, the input of PLC is connected with each photoelectric sensor and position sensor. And the output is directly connected with the driver of the transverse and longitudinal step motor to drive the operation of the stepping motor, so as to realize the seedling identification and orderly supply of the seedling. (1) Selection of PLC for programmable logic control PLC is responsible for collecting, analyzing and processing the output signal of the sensor in the system, and the output control signal is used to control the stepping motor to perform the corresponding action, which should be selected according to the number of I / O pins. The system requires at least 7 input points and at least 4 output points. According to the flexibility of programming and program debugging, expansion performance and performance-to-price ratio, the DVP14SS211T model of Delta Electronics is selected. It has 4 KB storage and 8 bit input and 6 pins output, two high speed pulse output terminals, built-in RS232 and RS485 communication port, can be connected with computer and LED
scrreen, and it eaasy to program m and debug. (2) Installaation of steppping motor an nd sensor Considerinng the need of differentt output torqque when trray conveyin ng mechanism m is used transversely t and lon ngitudinally, tthe 110BYG350CH-SAK KSMA-0501 ttype and 86B BYG350CH-S SAKSHL-03001 type threee-phase steppping mo otor produced by Hutchhison Compaany are usedd to drive the t transverse and longiitudinal shiftting mechannism resspectively. Thhere are two types t of senssor types usedd by the system. One is th he photoelecttric sensor ussed to detect and ideentify the seeedling. (4 BG GS-S08N back kground inhibbitory diffusee reflectance photosensorss produced by b OPTEX). The T oth her is a positiion sensor ussed to detect the position (the claw and the initial position p of thhe seedling mechanism). m T The req quired positioon sensor is innductor, whicch only deteccts metal objeects. Accordin ng to the sizee of the measu ured surface and thee required sennsing distance, we choosee the E2E-Z-M M10 type po osition sensor produced byy OMRON. According A to the fun nction of the stepping mootor and the detecting poosition requirred by the seensors, the sttep motor is installed on the transplanting deevice separateely, and the concrete distriibution is sho own in figure 10.
Fig g. 10
Installattion of sensorss and stepping motors
4.2 2.2 software ddesign (1) Control program floow Accordingg to the actionn requiremen nts of the inttelligent transsplanting con ntrol system, the flow chart of the onntrol sysstem for seeddling recognittion is proposed, and the coontrol system m program is composed, c ass shown in Fiigure 11.
Start system initialization Yes
No
Sensor 1 trigger? Yes Waiting for picking
Is the seedling supplying mechanism In the left limit? No
Yes Drive step motor to move longitudinally
Supplying Mechanism moves to the left limit
Tray moves longitudinalLy by 是 否 one grid
Drive step motor to move longitudinally Move the tray longitudinally to the picking position（tray distance switch signal detected）
Sensor 4 trigger? Yes
No
Exit sensor trigger？
Drive step motor to move horizontally
No
Sensor 3 trigger? Yes Waiting for picking
No
Exit sensor trigger？
Waiting for picking
No
Yes Exit sensor trigger？
No
Drive step motor to Yes move longitudinally
Yes
Tray moves to the left by one grid
Drive step motor to move horizontally Yes Tray moves to the right by one grid Yes
No Sensor 2 triggers?
Sensor 1 trigger?
Seven counts? Yes
No No Seven counts?
No
Sensor 2 triggers? Yes
Yes
No
Waiting for picking
Exit sensor trigger？
No
Yes No 15 counts?
Tray moves longitudinally by one grid
Drive step motor to move longitudinally
Yes End of the process
Fig. 11
Flow chart of control program for automatic identification of seedling
(2) Control system Circuit and Software Interface According to the program flow of the control system, the circuit diagram of the control system is designed, as shown in Figure 12, the I / O port of PLC is defined in Table 2. According to the circuit diagram of the control system, the man-machine interactive interface of the control system is designed (as shown in Figure 13), where two modes of operation (manual and automatic) can be switched, as well as pot distance, tray distance, number of rows of seedling tray, etc.
Fig g. 12
Circuit Diagram of Coontrol system for f automatic rrecognition an nd picking Seed dling
Table 2
I / O port definiition of PLC
Input
Output
X0
Leftmost lim mit (transverse) ssignal
Y0
Longitudinal step motor pulsse
X1
Seedling g tray limit signa nal
Y1
Longitudinal step motor pulsse
X2
Sk kip single
Y2
Transverse sstep motor pulsee
X3
Claw w exit signal
Y3
Transverse sstep motor pulsee
X4
X5
X6
X7
Signal 1 of riight shift detecttion of photoeelectric sensor Signal 2 of left l shift detectiion of photoeelectric sensor Signal 4 of ph hotoelectric sennsor for detection of th he right end of tthe tray Signal 3 of ph hotoelectric sennsor for detection of th he left end of thhe tray
a
Fig g.13
b
Control system interfaace
a Graphical G user innterface b Paraameter setting
5 Experimeent In order too test the quaality of the deesigned intellligent transpllanting system m, a comparaative experim ment was carrried outt in the laboraatory. The expperiment setu up was shownn in Figure 14 4, and the ind door illuminat ation was suffficient.
Fig g. 14 Indoor traansplanting exxperiment of seeedling identifiication
5.1 1 Condition n of experim ment In this expperiment, seeddlings at age of o 35d were uused for testin ng and the rate of emergennce of the seeedling was abbout 92% %. The speciific morpholoogical parameeters can be ffound in Secction 2.1. For statistics andd comparison n, the numberr of seeedlings in eacch tray was controlled c at 20 2 plants (thaat is, to contrrol the emerg gence rate to 85% for 128 8 pots). After the
tray was put into the conveying mechanism, the motor was started at 18.8Hz frequency to drive the continuous operation of the seedling picking mechanism and the rotary planting mechanism, corresponding to the transplanting speed of 90 plants /min. The whole experiment was divided into two groups. One group of control system did not turn on the function of automatic recognition and the other group turned on the function of automatic recognition. Each experiment was carried out until 10 trays were transplanted and the corresponding data were collected. 5.2
Test index
Important indicators to evaluate the performance of the transplanting system include the success rate of seedling acquisition and the rate of leakage. The definitions are as follows:
P=
T 100% T0
Eq. (1)
Where P is success rate of seedling acquisition; T is number of seedling acquisition; T0 is number of total seedlings.
D=
S 100% S0
Eq. (2)
Where D is leakage rate; S is number of dropping times; S0 is number of picking times. 5.3
Result of test
At the transplanting speed of 90 plants / min, the results of two groups with and without transplanting and identifying functions are shown in Table 3. In the Table 3, the number of seedling picking refers to the number of times that the seedling claw enters the pot during the process of seedling conveying. The number of successful seedlings picking is the number of seedlings that are successfully picked out from the pot with not damage (leaves and stems of seedlings have not been destroyed and more than 2 / 3 of the soil base is taken out). The number of successful seedlings dropping is the number of the seedling that are dropped into the planter with little damage (the stem and leaves of the seedlings are not broken and the soil base is not broken when contacting with the planter). Table 3 Results of transplanting experiments of two groups with and without the seedling recognition function Number
number of
The number
Number of
of
successful
of successful
seedlings
seedling
seedlings
seedlings
Transplanting
Successful
Working speed/plants· condition
min-1
Group 1： without the seedling
Tray No.
function
picking
picking
dropping
1
108
128
94
92
87.04
28.13
2
108
128
95
89
87.96
30.47
3
108
128
97
95
89.81
25.78
4
108
128
94
90
87.04
29.69
5
108
128
96
92
88.89
28.13
6
108
128
95
93
87.96
27.34
7
108
128
97
94
89.81
26.56
8
108
128
96
93
88.89
27.34
9
108
128
95
89
87.96
30.47
10
108
128
96
91
88.89
28.91
88.42
28.28
average value Group 2：with
recognition
rate/% picking
90
recognition
the seedling
Leakage rate of
90
1
108
111
98
89
90.74
19.82
2
108
107
93
90
86.11
15.89
3
108
105
94
90
87.04
14.29
function f
4
108
102
92
88
85.18 8
13.73
5
108
113
97
92
89.81
18.58
6
108
108
96
93
9 88.89
13.88
7
108
110
94
91
87.03 3
17.27
8
108
105
96
90
88.89 9
14.28
9
108
109
96
89
88.89 9
18.34
10
108
111
97
92
89.81
17.11
88.23 3
16.46
verage value av
5.4 4 Analysis aand discussioon The successs rate and leeakage rate in n two transpllanting statess were compaared and anallyzed when the t transplantting speeed was 90 / m min, as show wn in Figure 15.
a
Fig g. 15
b
Comparrison of transp planting resultss between two groups at diffeerent states wh hen transplantiing speed is 90 0 plants / min
a Picking P rate b Leakage rate
Accordingg to Fig. 15, under the co ondition of trransplanting speed s of 90 plants / min,, the automatic transplantting opeeration of the first group without the identificatioon function of the vacancy of seedlinggs was similaar to that of the intelligent transsplanting opeeration of the second grouup with the fu unction, and the success rrate of seedlings picking was w clo ose to that of the group 2. The averagee values of piicking rate were w 88.42% and a 88.23%, respectively. But the leakkage rate of the secoond group waas significanttly lower thann that of the first group, the t average rrate of the seecond group was w 28.28% and thee rate of the first f group waas 16.46%. T The main reasson is that the number of seedling rem moval is reducced, thaat is, the num mber of grabbiing soil base without seeddling is reduced, and the number n of tran ansplanting em mpty soil basse is red duced, whichh results in thhe decreased leakage ratee. Therefore, the designed d intelligent ttransplanting system can not onlly meet the reequirements of o high efficiiency transplaanting seedlin ng quality, bu ut also effectiively reduce the t phenomennon of seedling leakkage caused by b vacancy off seedlings; thhe leakage ratte is reduced by about 12% %. In the couurse of transsplanting exp periment, it is found thaat there are two kinds oof failure caases of seedlling ideentification: 1. There iss no seedling in the pot bu ut before the seedling claw w moves into the hole, thee seedling traay conveyer does d nott remove trayy in correct position, p whicch leads to thhe failure grab bbing. The main m reason fo for this situatiion is the sennsor misjudgment caaused by the interference i of o neighborinng seedlings. 2. There arre seedlings in i the pot but before the cllaws enter thee hole, the seedling tray coonveyer mov ves away the pot, p wh hich results inn the miss of grabbing g thiss seedling. Thhe main reaso on for this pheenomenon is that the seedling is entanggled witth neighborinng seedling or o the parts of o the machinne, which caauses the sen nsor to be unnable to identtify the seedlling
because of the sensitivity limitation of the sensors. Table 3 shows that the number of seedling picking is less than the total number of seedlings. Although this phenomenon had no effect on the leakage rate of the system, it caused a waste problem and affected the economy of seedlings transplanting. In view of the above problems, we can improve the design by adjusting the sensitivity of the photoelectric sensor and the tilt angle of installation, controlling suitable moisture content of the potted seedling (water content directly affects the upright degree of potted seedling during operation), removing the parts (pressure disc rod, seedling leaf stem) that may be wound with the leaf stem of the seedling on the condition of not affecting the working function of the mechanism, etc. Measures can be taken to improve the accuracy of automatic seedling identification, and reduce the rate of seedling leakage and further control the level of seedling waste.
6
Conclusions
(1) An intelligent transplanting system was proposed on the basis of the transplanting device for potted pepper seedlings. The system is composed of seedling picking mechanism with the combination five-bar mechanism and fixed axle train, seedling tray conveying mechanism, rotary planting mechanism, photoelectric sensor for seedling detection, position sensor, stepper motor and PLC control system. By the use of computer logic programming, the motor drives each mechanism in a coordinated and orderly manner, and it can detect whether the soil base with or without a seedling on it, and realize the automatic recognition of the seedling. (2) A comparative experiment on transplanting for pepper seedlings in laboratory showed that: when the transplanting speed is 90 plants/min, the intelligent system can effectively identify the condition of the soil base, so as to avoid picking the defect soil base and reduced leakage rate caused by the vacancy of seedlings; The success rate of seedling picking was 88.23% and the leakage rate was 16.46%. Compared with the transplanting model without seedling identification, the average rate of leakage was reduced about 12%.
Acknowledgements The work was sponsored by the National Key Research and Development Program of China Sub-project (No.2016YFD0700103 & No.2017YFD0700800), and the Innovation Scientists and Technicians Troop Construction Projects of Henan Province (No.184200510017).
[References] [1]
[2]
[3]
[4]
[5]
[6]
Wu J M., Zhang X C., Jin X., et al.: Design and experiment on transplanter pot seedling disk conveying and positioning control system. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE). 31(1), 47-52 (2015). (in Chinese with English abstract) Jin X., Du X W., Ji J T., He Z T., et al.: Automatic detection and controlling system of pot seedlings for transplanters. J. International Agricultural Engineering Journal. 24(2), 143-151 (2015). Yang Z Y., Zhang W Q., Li W., et al.: Monocular vision-based method for direction adjustment of transplanting potted-seedling leaves. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE). 30(14), 26-33(2014). (in Chinese with English abstract) Tai Y W.，Ling P P.，Ting K C.: Machine vision assisted robotic seedling transplanting. J．Transactions of the ASAE. 37(2), 661-667 (1994)． K H Ryu.，G Kim.，J S Han.: Development of a robotic transplanter for bedding plants. J．Journal of Agricultural Engineering Research. 78(2)，141-146 (2001)． Humphries S.，Simonton W.: Identification of plant parts using color and geometric image data. J．Transactions of the ASAE. 36(5), 1493-1500 (1993)．
[7] [8] [9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
Mizuochi Y，Dohi M.: Machine vision for transplanter of vegetables. C．ASAE Annual Meeting，2003． Albertus J V V.: Multiple Transplanting Apparatus[P]．US，No．20040020110A1，(2004)． HU F.，YIN W Q.，CHEN C R.，XU B.: Recognition and localization of plug seedling based on machine vision. J. Journal of Northwest A＆F University(Nat．Sci．Ed． 41(5), 183-188 (2013). Hou S.: Research on Methodology to identify Seedling Health Condition of Plug Seedling Transplanting Robot. Master Thesis, Jiangsu University, (2015). Yang Z Y., Zhang W Q., Li W., et al.: Information acquisition method of potted-seedling transplanting fitness using monocular vision. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 30(3), 112－119 (2014). (in Chinese with English abstract) He L Y., Yang T W., Wu C Y.: Optimization of Replugging Tour Planning Based on Greedy Genetic Algorithm. J. Transactions of the Chinese Society for Agricultural Machinery. 48(5), 36-43 (2017). (in Chinese with English abstract) Liu J D., Cao W B., Xu H Z., et al.: Adaptive fuzzy-PID control of accurate orientation for auto-detect seedling supply device. J. Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE), 33(9), 37－44 (2017). (in Chinese with English abstract) Jin X., Du X W., Ji J T., et al.: Physical characteristics of plug seedling transplanted by the return-blank type transplanter. J. International Agricultural Engineering Journal, 24(4), 1-10 (2015). Wang Y Y., Yu H Y., Liu Y Y.: Mechanical transplanting plug tray localization method based on binocular stereo vision. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 32(5), 43-49 (2016). (in Chinese with English abstract) Ni Y L., Jin C Q., Liu J.: Design and experiment of system for picking up and delivering seedlings in automatic transplanter. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE). 31(23), 10-19 (2015). (in Chinese with English abstract) Sun G X.: Automatic plug seedlings Transplanting Robot Based On Machine Vision. Master Thesis, Nanjing Agricultural University. (2009) Tan M H.，Cao Q X.，Qiu Q.，et al.: Real-time Localization Research of Transplanting Based on Machine Vision. J. Transactions of the Chinese Society for Agricultural Machinery. 47(2), 59-63 (2016). Jin X., Du X W., Ji J T., et al.: Mechanical property experiment of plug seeding with pots gripping-picking. J. International Agricultural Engineering Journal, 24(4), 24-33 (2015). Feng Q C., Wang X., Jiang K., et al.: Design and test of key parts on automatic transplanter for flower seedling. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 29(6), 21-27 (2013). (in Chinese with English abstract) Yang Z Y., Zhang W Q., Li W., et al.: Monocular vision-based method for direction adjustment of transplanting potted-seedling leaves. J. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE). 30(14), 26-33 (2014). (in Chinese with English abstract). Yang Y., Gao Q X., Sheng G D.，et al.: Plug Tray Localization and Detection System Based on Machine Vision. J. Transactions of the Chinese Society for Agricultural Machinery. 44(6), 232-235 (2013). (in Chinese with English abstract) Lu C H.，Chen X G., Wu W F.，et al.: Automatic Measuring System of Seedling Perpendicularity Based on Binocular Stereo Vision. J. Transactions of the CASE. 33(1), 60-62 (2002). WANG Y Y.，YU Y.，LIU Y Y.: Recognition of seedlings in mechanical transplanting processing by Morphological image detection. J. Optics and Precision Engineering. 24(3), 651-658 (2016). FENG L., CHAI R Y.，SUN G M.，et al.: Identification and Classification of Rice Leaf Blast Based on Multi--Spectral Imaging Sensor. J. Spectroscopy and Spectral Analysis. 29(10), 2730-2733 (2009). Deng W., Huang Y B., Zhao C J., Wang X.: Identification of seedling cabbages and weeds using hyperspectral
imaging. J. Int J Agric & Biol Eng. 8(5), 65-72（2015）. [27] Narendra V G., Hareesh K S.: Study and comparison of various image edge detection techniques used in quality
[28]
[29]
[30]
[31]
inspection and evaluation of agricultural and food products by computer vision. J. Int J Agric & Biol Eng. 4(2), 83-90 (2011). Wu Q., Wang C., Fang J J., Ji J W.: Field monitoring of wheat seedling stage with hyperspectral imaging. J. Int J Agric & Biol Eng. 9(5), 143-148 (2016). Wang Q H., Tang Y H., Xiao Z.: Grape size detection and online gradation based on machine vision. J. Int J Agric & Biol Eng. 10(1), 226-233 (2017). Li B R., Long Y., Song H B.: Detection of green apples in natural scenes based on saliency theory and Gaussian curve fitting. J. Int J Agric & Biol Eng. 11(1), 192-198 (2018). Jin X., P J., Ji J T., et al.: Experiment and simulation analysis on high-speed up-film transplanting mechanism. J. International Agricultural Engineering Journal, 26(3), 105-112 (2017).
Jin Xin, PhD, Associate Professor, College of Agricultural Equipment Engineering, Henan University of Science and Technology. Dr. Jin focuses on application of computer programming and sensor Technology in Agriculture.
Zhao Kaixuan, PhD, Lecturer, College of Agricultural Equipment Engineering, Henan University of Science and Technology. Dr. Zhao focuses on artificial intelligence and image processing algorithm. Ji Jiangtao, PhD, Professor, College of Agricultural Equipment Engineering, Henan University of Science and Technology. Prof. Ji focuses on the study of intelligent and information technology of agricultural machinery equipment. Du Xinwu, PhD, Associate Professor, College of Agricultural Equipment Engineering, Henan University of Science and Technology. Dr. Du focuses on design of agricultural machinery and analysis of field experiment Ma Hao, PhD, Lecturer, College of Agricultural Equipment Engineering, Henan University of Science and Technology. Dr. Ma focuses on application of big data analysis and information processing technology in agriculture. Qiu Zhaomei, PhD, Associate Professor, College of Agricultural Equipment Engineering, Henan University of Science and Technology. She focuses on the study of computer
programming and control system.
Jin Xin
Zhao Kaixuan
Ji Jiangtao o
Du Xin nwu
Ma Hao
Qiu Zhaomei Z
Research Highlights: 1. Propose the idea of intelligent transplanting in field planting of agricultural production. 2. On the basis of computer logic programming, combined crop seeding, photoelectric sensor and mechanical system, we developed an intelligent control system to realize the active recognition and effective collection of seedling in the transplanting process. 3. The average rate of missed planting of the machine with intelligent control system was reduced about 12%, which has the potential to solve the miss planting problem caused by vacancy of seedlings.