- Email: [email protected]
A computer-assisted system for large-scale engineering mapping
Photogrammetria, 36 (1981) 203--216
203
Elsevier Scientific Publishing Company, Amsterdam -- Printed in The Netherlands
A COMPUTER-ASSISTED SYSTEM F O R LARGE-SCALE E N G I N E E R I N G MAPPING*
R.P. CLEAVES Australian Aerial Mapping Pty. Ltd. Windsor, Vic. (Australia)
(Received November 5, 1980; accepted February 10, 1981)
ABSTRACT Cleaves, R.P., 1981. A computer-assisted system for large-scale engineering mapping. Photogrammetria, 36: 203--216. A digital mapping system utilising existing analogue stereoplotting instruments, a minicomputer and microprocessors, which was developed under strict budget limitations over a three year period, is described. The rationale of the development, applications of the system and the effect on personnel are considered.
INTRODUCTION As th e c o m p l e x i t y of land m a n a g e m e n t and civil engineering problems have increased, particularly in the ur ba n and near-urban areas, engineers charged with the responsibility of finding solutions have naturally sought to use the latest t e c h n o l o g y available, with its inevitable use of the high-speed digital c o mp u ter . As a map or plan is basic t o finding these solutions, it is simple logic t hat there is or will be, an increasing dem a nd for maps and plans in digital form in the years ahead. Traditionally, the provision o f these maps and plans have been at large scales, sufficient for the project u n d e r consideration. T he large scales have enabled the p h o t o g r a m m e t r i s t t o plot in m u c h greater detail the man-made improvements on t he t o p o g r a p h y , as well as c o n t o u r s at m uch closer intervals. To p r o d u c e this digital data will inevitably involve mapping organisations in a capital investment in new t echnol ogy. The problem for small private organisations, with limited resources, is t o decide w h e t h e r to re-equip with sophisticated new i ns t r um ent at i on, such as the analytical plotter, or t o att e m p t to upgrade existing e q u i p m e n t to m e e t the new challenge. To q u o t e Leatherdale and Keir (1979), " e ach mapping organisation has to *Presented paper, 14th Congress of the International Society for Photograrnmetry, Hamburg 1980, Commission IV, Working Group I.
0031-8663/81/0000--0000/$02.50 © 1981 Elsevier Scientific Publishing Company I
!
204 react to new t e c h n o l o g y in its own way in relation t o its role for the fut ure and cu r r en t constraints of expertise, e q u i p m e n t and capital resources." Present-day analogue instruments f r om the major manufacturers currently in use are m o r e t ha n adequate in terms o f their precision for the task of providing digital map data. It would seem logical to upgrade this e q u i p m e n t with the addition of electronic c o m p o n e n t s to acquire digital data as an alternative to, or during the process of, obtaining conventional line plotting. This is n o t to say t h a t analytical plotters do n o t have a role to fill, but simply t h a t if the objective is digital topographic data, the sophistication and expense o f new i n s t r u m e n t a t i o n is hard to justify in terms of p r o d u c t cost. This paper describes a digital mapping system which has been built up by stages over the past three years, utilising existing analogue instruments and m i n i c o m p u t e r which have been upgraded to a full system with the aid of microprocessors. HARDWARE CONFIGURATION In 1976 o u r c o m p a n y was equipped with eleven analogue plotting instruments, including one capable of producing o r t h o p h o t o s , and its own minic o mp u ter . This e q u i p m e n t had been purchased over the previous sixteen years. Th e m i n i c o m p u t e r was one of the later acquisitions, having been purchased some t w o years previously at a time when spending on timesharing bureaux had reached alarming proportions. It was also becoming increasingly obvious that m a n y of the c o m p a n y ' s clients were awakening t o the potential o f the high-speed c o m p u t e r . In considering the hardware for a computer-assisted system in the private enterprise environment, it was necessary to consider the main types of work to be served as well as the capital investment required. In the main we, as a private organisation, have no clearly defined task in hand and are o f t e n called u p o n to p e r f o r m in widely diversified areas. (Cleaves, 1978). Although the prime task would be large-scale mapping, it was considered unwise to exclude small-scale topographic mapping, just as it would be unwise n o t to consider the physical location f r om which source data would be obtained. If all source data was to be obtained at the same location then a total " o n lin e" system could have been considered. However, as the physical location f r o m which some of the source data would be obtained was m any miles away, it was vital that b o t h " o n l i ne " and " o f f l i n e " systems be considered. It soon became apparent t hat an " o n l i ne" system would need a central processor o f considerable power, in the order of 100K core m e m o r y at least, b u t t h a t any b r ea kdow n, even a m i n o r br e a k dow n, could bring the entire digital system to a halt. F o r p r o d u c t i o n reasons it was decided this was unacceptable; the risk of " d o w n " time on the central processor being all t o o evident. T he capital investment on a pow e r f ul central processor was also quite substantial when
205
coupled with the equipment necessary for acquisition, editing and presentation of results. For reasons of capital investment as well as a desire not to be dependent on a single central processor, an "off line" system based on microprocessors at the various input and edit stations, interfaced to a minicomputer of lesser core memory, was designed. The design had a number of advantages, one being that it could be built up by stages as required, thereby reducing the demands on the limited capital available. The microprocessor is a powerful tool with the advantage that as it is programmable it is possible to define the prime task, but retain the flexibility to modify and improve procedures as expertise increases. In terms of capital investment it was not necessary to commit the company to expensive equipment which could prove to be unsatisfactory, or to equipment which had been designed for specific tasks only. The hardware configuration as shown in Table I comprises acquisition, processing, editing and presentation equipment, which is discussed in greater detail in the following paragraphs. TABLE I
A. A. M. DIGITAL DATA ACQUISTION
L ~
2~KM,CRO L.~ PROCESSOR
[MASTER L l C°NS°LE ]
~STEREOPLCTTER __
m~-~ ISTEREO LD TER
KEYBOARD '
L
[DUAL DISKETTESI '
I i
I
SYSTEM PROCESSING
2LK MICRO I
P OCESSO
I
l
__,~ -j L
MU~T,PtRTxER
~
V,SUAL / DISPLAY ]
GRAPHIC ] SCREEN I
U.I
~ F~K,~BIT~Of
/Mml COMPUTERI
EDITING
I
ALPHA/NUMERIC
,S AL
DISPLAY
32K MINI -] COMPUTER~ 10 MEGAB¥rE I - L_?gc
2LKMICRO I I
~ I C ~
I I [
~3TTE~Rj
] ~ STEREOPLOTTER
~ ~
L.
~sq
~
24K MICRO PROCESSOR KEYBOARD DUAL DISKETTES PAPER TAPE PUNCH COORDINATE REGISTRATION UNIT PAPER TAPE PUNCH
o,sc UNIT
,jN r
]
I I
MAGNET,OTAP~ DN,T c,TRACK 800 BPI CARD READER PAPER TAPE READER
PRESENTATION
]
FLAT BED j PLOTTER
[
CONTROL~] INTERFACEJ
206
By installing microprocessors it is possible to interface a single stereoplotting instrument to a minicomputer, to develop software, to learn and understand the problems of capturing (?) digital data and to acquire the necessary expertise, all at a relatively low capital cost.
Data acquisition The equipment currently capable of digital data acquisition consists of a Wild A10, t w o Wild A8s and a Wild B8, together with a recently acquired Summagraphic digitizer table which can be used to digitize existing plans, or to assist the editing function. All of the stereoplotters were originally purchased as analogue plotting instruments for conventional line plotting and with the exception of the A10 have been fitted with 24K microprocessors which can be operated either " o f f line" as stand-alone equipment, or "on line" interfaced directly to the minicomputer. The A10, when originally purchased, was fitted with the EK8 for automatic coordinate registration to provide punched paper tape for aerial triangulation computations. To date the A10 has not been u p g r a d e d ; p a r t l y because the EK8, although an unintelligent device, is capable of digitizing " o n the fly" and thus can be used to obtain limited digital data. The A10 and one of the A8s are located many hundreds of miles away from the central processor. Both instruments are fitted with ahigh-speed paper tape punch which provides the means of transporting high-volume digital data over long distances. Although it is feasible to interface these instruments to the central processor through a telephone land line, this is considered to be uneconomic at the present time. Two of the instruments, one A8 (Fig. 1) and the B8, together with the graphic digitizer table (Fig. 2), which are located with, and interfaced directly to, the central processor via their microprocessors, are also equipped with
Fig. 1. Analogue Stereoptotter with 24 K Microprocessor.
Fig. 2. Graphic digitizer with microprocessor.
207
video screens which have a switch capability enabling them to be used as either alpha/numeric terminals or as graphic screens, using a 256 square dot matrix. These microprocessors are programmed to sample through the u p / d o w n counters, the square wave impulses from the shaft encoders, or in the case of the graphic digitizer table the inbuilt microprocessor, on a continuous basis, so long as the foot switch of the equipment is depressed. In theory t h e y are capable of sampling at the rate of 2400 characters per second, but in practice this is reduced by time filtering to approximately 20--25 per second. The registrations are retained in the machine system, that is a local system of coordinates, and are c o m p u t e d and displayed on the video screen when used as a graphic screen. As each planimetric feature or complete c o n t o u r is digitized, it is displayed and the operator has the ability to either reject or save the data according to his assessment of its correctness. If he chooses to save the data, it is stored on a permanent file on one of the diskettes. On completion of the digitizing of model or plan, the data is transferred to the minicomputer. The microprocessors can be operated in either of two modes, one mode is " t r a n s p a r e n t " where the microprocessor behaves essentially as a non-intelligent Visual Display Unit, the other mode is intelligent where the microprocessor responds to and transmits data to the minicomputer. In its " t r a n s p a r e n t " mode the microprocessor (Fig. 3) is "signed o n " to the minicomputer, enabling it to be identified. In this mode the microprocessor is under the control of the minicomputer, which establishes and allocates file space.
Fig. 3. Microprocessor Videoscreen and Disks. Fig. 4 . 4 8 K Mini Computer.
208 Once the microprocessor has been "signed o n " it reverts to its second mode, that of an intelligent unit, acting together with the minicomputer, and begins to transmit the stored data. The data is transformed b y the minicomputer into the terrain system of coordinates, a mathematical filtering of the data, based on a simple test of the deflection angle at the third point in any trio, is applied, and the data is stored on the 10 megabyte disc.
Processing The processing equipment consists of a Data General Nova 840 (Fig. 4), which was enhanced with various add-on units to enable the storage of highvolume digital data. The c o m p u t e r is m e m o r y - m a p p e d and can be operated foreground/background in time-shared BASIC and F O R T R A N dedicated to a single application. It is disc based and supports in-house "time-sharing" using BASIC thus the full core m e m o r y available in BASIC is available to each user. There are no hard and fast rules applying to operation of the system, b u t usually it is operated in time-shared BASIC during the day, with batch F O R T R A N run overnight. During the daytime, processing the data-acquisition stage is given priority over other users when data is to be transferred from the microprocessors. The system is used for all computations, including commercial processing, j o b costing, payroll, management reports etc., aerial triangulation computations, volume capacity and earthwork calculations. The minicomputer also provides the drive to the flatbed plotter through use of the spooling device. The plot instructions are c o m p u t e d in Fortran overnight, the data having been set up in BASIC during the previous daytime processing. Usually the plotter is left on overnight and commences plotting during the night until it needs operator intervention. The remainder of the plot instructions are left in the spool and provide the data to keep the plotter going during the next day's processing.
Editing Editing of the digital data before any hard-copy plotting is attempted is in all probability the most vital stage in a digital system. The equipment utilised here consists of a Tektronix 4010 storage display screen (Fig. 5) connected to a Nova 4, 32K minicomputer which provides the computing power. Data to be displayed and edited is stored on a 10 megabyte disc which is interchangeable with the 10 megabyte disc on the central processor. The digital data is displayed on the screen and edited by use of the cross hairs. Software has been developed to enable the data to be windowed, deleted, moved, added to, displayed at different scales, automatically squared, contours to be joined and so on. The procedure adopted has been to digitize data either on the stereoplotter or the graphic digitizer table and display it on the video screen, using the
209
Fig. 5. Storage Display Screen.
Fig. 6. Hybrid Flatbed Plotter.
raster scan d o t matrix. The data is transferred to the minicomputer, transformed into terrain units and re-displayed on the storage screen where final editing is completed prior to hard-copy plotting. Presentation
At the present time, flatbed plotting is achieved by using a hybrid plotter (Fig. 6) consisting of an analogue instrument plotting table which has been modified by the addition of stepping motors to the x and y drive spindles (Fig. 7) and interfaced to the minicomputer. The equipment does not have the quality of line work produced by a recognised flatbed plotter; however, this is often unimportant in civil engineer° ing work provided it is capable of maintaining required accuracies. Final cosmetics are still obtained in the conventional way. Software routines have been developed for a wide range of graphic output, from edited and reduced detail or c o n t o u r plots to capacity curves, c o n t o u r generation from random input, gridded control sheets, point plots of field generated "string" data, long sections, cross sections and so on. Ultimately it is intended to install a high-speed flatbed plotter which will enable plans to be produced with a higher presentation quality, so that with the minimum of additional work they can be reproduced and delivered to the client.
210
Fig. 7. Flatbed Plotter Control Interface.
Cost effectiveness Most if not all, private mapping concerns are faced with a restricted capital budget and our c o m p a n y was no exception. The hardware capital cost of the digital system, excluding the original cost of the stereoplotters, has been $A160,000 which is comparable with the price quoted in Australia for one analytical plotter, and this a m o u n t has enabled four analogue stereoplotters to be integrated into the system. Before purchasing equipment, cost effectiveness has to be carefully scrutinised. To be cost effective in the commercial world all equipment purchased must be utilised to the m a x i m u m and must improve upon the efficiency or accuracy of a current procedure. It must also produce a marketable product which is otherwise unobtainable, or a product at least comparable with and at lower cost than existing products (Cleaves, 1978). It is very difficult to assess the cost effectiveness of a digital system. Contrary to many opinions, digital data is not obtained almost as a by-product of the plotting procedure. In fact, the cost of producing digital data can be almost 100% on the cost of producing a graphical result. The true cost effectiveness or advantages are not experienced at the acquisition stage but rather at the user stage. We have found that stereoplotting time for man-made features has increased by up to 25% and that plotting of contours and drainage increases plotting time by 15%. Onto these increased times must be added the time taken to edit the data, which we have found to be around 40% to 50% of the plotting time, including the time needed to edge match model joins and sheet edges. Besides the increase in time necessary to produce a digital map, there is the additional capital investment in the equipment necessary to produce this data, which has to be offset against the cost of the product. However, if the p r o d u c t being produced is a new product, providing data that will be analysed m a n y times by the introduction of new parameters,
211
then it can certainly be considered to be cost effective, particularly if the new parameter permutations are innumerable. There are some cost savings to be made by a mapping organisation with a digital system. These savings are mostly seen in the areas of aerial triangulation and the c o m p u t a t i o n of volumes where the digital system enables rapid on-the-spot checks to be made on data which is otherwise hard to check, or which, when checked, involves costly re-observations. PERFORMANCE, APPLICATIONS AND S T A F F SATISFACTION
As with any new technology it is necessary to evaluate performance against expectation. The expectations for the digital system were threefold: (1) aerial triangulation; (2) engineering c o m p u t a t i o n s ; a n d (3) digital map production. Prior to establishing the hardware necessary to interface analogue instruments to a computer, considerable expertise had been built up in aerial triangulation and engineering computations using batch processing and timesharing bureaux. The transition from external computers to an "in-house" minicomputer had already been made and much of the expertise built up with outside computers was transferred and enhanced with the company's own computer.
Aerial triangulation Computation procedures for strip and two-dimensional block adjustments were firmly established. After purchase of the minicomputer a three-dimensional block adjustment program was developed, based on the xy program developed at Melbourne University by Bervoets (1971). Using a microprocessor " o n line" to the computer, the assembly of models into strips takes place as each model is read, with residuals on the pass points displayed on the video screen. Operator job satisfaction is enhanced and job costs reduced as less mistakes and re-observing of models have occurred. Procedures have now been developed which retain the digital data in the system from the time of original observation until gridded control sheets are returned to the instruments for the scaling of individual models. The'operators are supplied with elements of orientation, together with a model diagram which displays all points to be found on the model in their approximate position, plus the residuals of that model with respect to the final mean values. Flatbed-plotted control sheets for all plans on the job are provided, each sheet having a punch register position plotted which enables the sheets to be accurately joined on a stud register to all the surrounding sheets, thus removing all the problems of edge matching of sheets both on the plotting table and in the drawing office. If required, punch registered overlays for
212
all control sheets can also be provided, with the c o m p u t e r accurately controlling the punch register positions. J o b satisfaction has been considerably enhanced for almost of personnel as a result, with ready acceptance of the benefits that the system provides. Cost savings have also been measured, both in the reduction of personnel directly involved plus an increase of throughput as a result of the increased accuracy and reliability of the products produced during this phase of plan production.
Engineering computations For some years prior to the installation of the digital system, programs had been developed to c o m p u t e volumes, areas, capacity diagrams and earthwork quantities. The data preparation for these types of calculations had been either b y hand or b y point m o d e registration units onto punched paper tape or punched cards, with processing carried o u t on c o m p u t e r time-sharing bureaux. With the installation of the microprocessors " o n line" to the computer, the observation and recording of input data was greatly simplified and checking of the complex shapes is overcome b y displaying the profiles on the video screen at the acquisition stage. In simple cases where one of the surfaces is a plane, such as a reservoir capacity, it is also possible to c o m p u t e volumes and areas as the observation of the data is proceeding, such that there is no time delay between completion of observations and achieving a final result. Hard c o p y of these results is easily obtained b y storing progressive figures in a file in the c o m p u t e r and printing the results on the line printer. The methods of digitizing, joining of two irregular surfaces and computation of volume has been adequately described by Leatherdale and Keir (1979). However, for a number of reasons clients still require their results in terms of horizontal profiles, as represented b y contours, rather than vertical profiles which clearly define the complex shapes found in quarries or open-cut mining. A method of translating data digitized in vertical profiles into horizontal profiles has been developed, which enables these complex shapes to be digitized in the most accurate and economical way and yet satisfy client requirements for area and volumes to be presented in a form most usable by them. Over the past few years all of the major reservoirs which have either been investigated or built in the vicinity of Melbourne, a city of a b o u t three million people, have been surveyed by use of digital photogrammetry. Some five major reservoirs have been, or are being built. Initial capacity tables and capacity diagrams were prepared during the investigation stage, earthwork volumes were c o m p u t e d for the dam wall construction, b o t h the material removed in excavation to bedrock and the materials placed in the wall construction. On completion, as-built plans and final capacity tables and capacity curve diagrams have been prepared.
213
A particular problem on which the digital system is used has been the monitoring of quarries, both during their excavation and afterwards, when they are generally taken over by local councils, who use them as rubbish dumps until refilled. They are then landscaped and returned to the comm u n i t y as parkland or recreational areas. Although an apparently mundane problem, considerable sums of m o n e y are involved as well as safety and public health risks where clay or sand pits are close to, or now within, the suburban sprawl of a major city. The digital system is a fast, economical way of monitoring these problems and enables accurate volume and area figures to be produced which engineers use to specify rate of fill, final desired surface as well as reserves, in terms of remaining space and so forth.
Digital map production Digital map production is the last of the three objectives of the digital system and could n o t be attempted until a successful interface between stereoplotting instruments and computer had been achieved. Development of the procedure, both hardware components and the considerable software requirements, has dominated the activities of the computer room staff. The digital map production procedure has been developed around a modified feature code based on a recommended standard of feature codes developed by the National Mapping Council of Australia (Standards Association of Australia, 1980}. The modified feature code of four digits can be computerextended to the eight-digit feature code required by the standard. The microprocessors have been programmed to provide a p r o m p t facility in the form of a menu display, as an aid to the stereoplotter operators in selecting the feature codes to be used. The stereoscopic models can be digitized in
a
1'500
Fig. 8. Examples of digital mapping.
b
1"2500
214 any sequence and it is not necessary to maintain a particular feature code once selected, although for convenience this is usually attempted. The microprocessors are wired into the footswitch of the instruments so that they are activated as soon as the footswitch is depressed. The m e t h o d of operation is for the stereoplotter operator to key in the feature code and proceed to plot. The data recorded by the microprocessor consists of a pen command activated by the action of the footswitch and the XYZ coordinates of the instrument axis. In this way the data recorded is minimised, with sufficient information to enable identification by the c o m p u t e r at presentation stage. Data is separated into its respective feature code and colour code when it is transferred to the minicomputer. This facilitates editing and flatbed plotting, a~s line sizes and types, symbols etc. are matched up with particular feature or colour codes. The digital system has been used for mapping in the scale range 1:500 to 1:100,000 with magnetic tape supplied to clients for further use. An example of digital mapping is shown at Fig. 8 with 1:500 and 1:2,500 maps produced from the same data. The fundamental aim of this organisation has been to upgrade existing analogue instruments by the addition of microprocessors and related peripherals to acquire digital map data in the most economical way. The minimum data necessary to describe any line is captured by sampling at a high rate, and mathematically reducing this data such that the plot produced from the reduced digital data would faithfully reproduce a plot produced from the table of a stereoplotting instrument. CONCLUSION In development of the digital system our aim has been one of staff involvement in new technology to produce a new product efficiently. Staff must obtain job satisfaction and must not feel their security is threatened by the instrumentation of new technology. It is very difficult to obtain cooperation and feedback from highly skilled stereoplotter operators by telling them that the new equipment can be operated by semi-skilled people. Equally, job satisfaction deteriorates rapidly if all so-called tedious tasks are removed only to be replaced by a lesser number of even more tedious tasks. There is no job satisfaction whatsoever to a highly skilled stereoplotter operator who is asked to spot height a stereo model on a tight grid mesh for the c o m p u t e r to produce contours. He would much rather produce the contours and acquire digits during the process. The fact that this does not lend itself to easy computer processing is of no interest, for computers are supposed to enhance his lifestyle. The social consequences of the new technology have to be considered if retrenchment of skilled staff, or lack of o p p o r t u n i t y to y o u n g eager graduates
215
of academic colleges is to be avoided. It is very difficult for a small c o m p a n y faced with an explosion in costs, particularly labour costs, to consider this last point, b u t if a successful transition to a digital system is to be attempted it has to be considered. After all, a digital system is complex and requires highly intelligent people to develop and operate it. POSTSCRIPT
This paper was originally written in January 1980 and published at the I.S.P. congress in Hamburg in July 1980. Since then, continuous research and development has naturally produced some changes to the system described. In addition, because of the development in microprocessor technology the author can see a major philosophical change in future digital mapping systems. The A.A.M. digital system is what is known as a distributed processing system. The main advantage of such a system is that it is n o t dependent on a single source of intelligence (i.e. the central processor), b u t has local intelligence via the microprocessors at the stereoplotters. Thus the entire system is not brought to a halt in the event of a hardware failure in the central processor. The initial development of this system called for 24K microprocessors, programable in the BASIC language, which were capable of acquiring digital data and transferring it to the central processor. The " o n line" processing p o w e r was limited to displaying the data on the videoscreen in the local coordinate system and to some simple edit routines. These microprocessors have n o w been upgraded to 64K bytes, which are programmable in F O R T R A N and which have a far greater " o n line" processing capability. As such they are n o w able to carry o u t processing tasks previously done by the central processor. Transformations from the local coordinate system of the stereoplotter to the terrain coordinate system are n o w carried o u t in real time at the acquisition stage, together with the real time display of the graphics on the videoscreen. In addition calculations for perspective centre coordinates, model join residuals, areas and volumes, and absolute and relative orientation are all carried o u t in real time b y the microprocessors at the stereoplotters. Thus if the central processor fails at any time due to a hardware fault the acquistion stations (stereoplotters or graphic digitizer tables) can continue to work, storing data on diskettes if required. The major philosophical change as seen b y this author is the disappearance of the central processor altogether from future digital mapping systems. Current microprocessor technology is such that all of the tasks carried o u t by the central processor can be distributed to a series of microprocessors interlinked to enable data to be transferred around the system as and when necessary. Thus the edit stations can be microprocessor supported, and the flatbed or automatic plotting table can be microprocessor controlled. A b o u t the only c o m p u t a t i o n which couldn't be handled b y a microprocessor at present would be a Bundle block adjustment.
216 The major advantage t o such a distributed system will be t h a t hardware failure to any part of the system will only affect t h a t part, with all o t h e r functions o f th e system operating normally. As microprocessors are relatively inexpensive it will p r o b a b l y be possible t o carry a c o m p l e t e spare system which can be plugged in t o any part of the system in the event of a hardware failure. REFERENCES Leatherdale, J.D. and Keir, K.M., 1979. Digital methods of map production. Photogramm. Rec., 9 (54): 757--772. Cleaves, R.P., 1978. Mini-Map (A semi-automated mapping system). 21st Austr. Surv. Congr., Adelaide, April, 1978, pp. 135--142. Bervoets, S.G., 1971. Horizontal block adjustment, Austr. Sure., 23 (5): 304--317. Standard for the exchange of feature coded digital mapping and charting data on Magnetic Tape. DR 79178 Standards Association of Australia, 1980. -
203
Elsevier Scientific Publishing Company, Amsterdam -- Printed in The Netherlands
A COMPUTER-ASSISTED SYSTEM F O R LARGE-SCALE E N G I N E E R I N G MAPPING*
R.P. CLEAVES Australian Aerial Mapping Pty. Ltd. Windsor, Vic. (Australia)
(Received November 5, 1980; accepted February 10, 1981)
ABSTRACT Cleaves, R.P., 1981. A computer-assisted system for large-scale engineering mapping. Photogrammetria, 36: 203--216. A digital mapping system utilising existing analogue stereoplotting instruments, a minicomputer and microprocessors, which was developed under strict budget limitations over a three year period, is described. The rationale of the development, applications of the system and the effect on personnel are considered.
INTRODUCTION As th e c o m p l e x i t y of land m a n a g e m e n t and civil engineering problems have increased, particularly in the ur ba n and near-urban areas, engineers charged with the responsibility of finding solutions have naturally sought to use the latest t e c h n o l o g y available, with its inevitable use of the high-speed digital c o mp u ter . As a map or plan is basic t o finding these solutions, it is simple logic t hat there is or will be, an increasing dem a nd for maps and plans in digital form in the years ahead. Traditionally, the provision o f these maps and plans have been at large scales, sufficient for the project u n d e r consideration. T he large scales have enabled the p h o t o g r a m m e t r i s t t o plot in m u c h greater detail the man-made improvements on t he t o p o g r a p h y , as well as c o n t o u r s at m uch closer intervals. To p r o d u c e this digital data will inevitably involve mapping organisations in a capital investment in new t echnol ogy. The problem for small private organisations, with limited resources, is t o decide w h e t h e r to re-equip with sophisticated new i ns t r um ent at i on, such as the analytical plotter, or t o att e m p t to upgrade existing e q u i p m e n t to m e e t the new challenge. To q u o t e Leatherdale and Keir (1979), " e ach mapping organisation has to *Presented paper, 14th Congress of the International Society for Photograrnmetry, Hamburg 1980, Commission IV, Working Group I.
0031-8663/81/0000--0000/$02.50 © 1981 Elsevier Scientific Publishing Company I
!
204 react to new t e c h n o l o g y in its own way in relation t o its role for the fut ure and cu r r en t constraints of expertise, e q u i p m e n t and capital resources." Present-day analogue instruments f r om the major manufacturers currently in use are m o r e t ha n adequate in terms o f their precision for the task of providing digital map data. It would seem logical to upgrade this e q u i p m e n t with the addition of electronic c o m p o n e n t s to acquire digital data as an alternative to, or during the process of, obtaining conventional line plotting. This is n o t to say t h a t analytical plotters do n o t have a role to fill, but simply t h a t if the objective is digital topographic data, the sophistication and expense o f new i n s t r u m e n t a t i o n is hard to justify in terms of p r o d u c t cost. This paper describes a digital mapping system which has been built up by stages over the past three years, utilising existing analogue instruments and m i n i c o m p u t e r which have been upgraded to a full system with the aid of microprocessors. HARDWARE CONFIGURATION In 1976 o u r c o m p a n y was equipped with eleven analogue plotting instruments, including one capable of producing o r t h o p h o t o s , and its own minic o mp u ter . This e q u i p m e n t had been purchased over the previous sixteen years. Th e m i n i c o m p u t e r was one of the later acquisitions, having been purchased some t w o years previously at a time when spending on timesharing bureaux had reached alarming proportions. It was also becoming increasingly obvious that m a n y of the c o m p a n y ' s clients were awakening t o the potential o f the high-speed c o m p u t e r . In considering the hardware for a computer-assisted system in the private enterprise environment, it was necessary to consider the main types of work to be served as well as the capital investment required. In the main we, as a private organisation, have no clearly defined task in hand and are o f t e n called u p o n to p e r f o r m in widely diversified areas. (Cleaves, 1978). Although the prime task would be large-scale mapping, it was considered unwise to exclude small-scale topographic mapping, just as it would be unwise n o t to consider the physical location f r om which source data would be obtained. If all source data was to be obtained at the same location then a total " o n lin e" system could have been considered. However, as the physical location f r o m which some of the source data would be obtained was m any miles away, it was vital that b o t h " o n l i ne " and " o f f l i n e " systems be considered. It soon became apparent t hat an " o n l i ne" system would need a central processor o f considerable power, in the order of 100K core m e m o r y at least, b u t t h a t any b r ea kdow n, even a m i n o r br e a k dow n, could bring the entire digital system to a halt. F o r p r o d u c t i o n reasons it was decided this was unacceptable; the risk of " d o w n " time on the central processor being all t o o evident. T he capital investment on a pow e r f ul central processor was also quite substantial when
205
coupled with the equipment necessary for acquisition, editing and presentation of results. For reasons of capital investment as well as a desire not to be dependent on a single central processor, an "off line" system based on microprocessors at the various input and edit stations, interfaced to a minicomputer of lesser core memory, was designed. The design had a number of advantages, one being that it could be built up by stages as required, thereby reducing the demands on the limited capital available. The microprocessor is a powerful tool with the advantage that as it is programmable it is possible to define the prime task, but retain the flexibility to modify and improve procedures as expertise increases. In terms of capital investment it was not necessary to commit the company to expensive equipment which could prove to be unsatisfactory, or to equipment which had been designed for specific tasks only. The hardware configuration as shown in Table I comprises acquisition, processing, editing and presentation equipment, which is discussed in greater detail in the following paragraphs. TABLE I
A. A. M. DIGITAL DATA ACQUISTION
L ~
2~KM,CRO L.~ PROCESSOR
[MASTER L l C°NS°LE ]
~STEREOPLCTTER __
m~-~ ISTEREO LD TER
KEYBOARD '
L
[DUAL DISKETTESI '
I i
I
SYSTEM PROCESSING
2LK MICRO I
P OCESSO
I
l
__,~ -j L
MU~T,PtRTxER
~
V,SUAL / DISPLAY ]
GRAPHIC ] SCREEN I
U.I
~ F~K,~BIT~Of
/Mml COMPUTERI
EDITING
I
ALPHA/NUMERIC
,S AL
DISPLAY
32K MINI -] COMPUTER~ 10 MEGAB¥rE I - L_?gc
2LKMICRO I I
~ I C ~
I I [
~3TTE~Rj
] ~ STEREOPLOTTER
~ ~
L.
~sq
~
24K MICRO PROCESSOR KEYBOARD DUAL DISKETTES PAPER TAPE PUNCH COORDINATE REGISTRATION UNIT PAPER TAPE PUNCH
o,sc UNIT
,jN r
]
I I
MAGNET,OTAP~ DN,T c,TRACK 800 BPI CARD READER PAPER TAPE READER
PRESENTATION
]
FLAT BED j PLOTTER
[
CONTROL~] INTERFACEJ
206
By installing microprocessors it is possible to interface a single stereoplotting instrument to a minicomputer, to develop software, to learn and understand the problems of capturing (?) digital data and to acquire the necessary expertise, all at a relatively low capital cost.
Data acquisition The equipment currently capable of digital data acquisition consists of a Wild A10, t w o Wild A8s and a Wild B8, together with a recently acquired Summagraphic digitizer table which can be used to digitize existing plans, or to assist the editing function. All of the stereoplotters were originally purchased as analogue plotting instruments for conventional line plotting and with the exception of the A10 have been fitted with 24K microprocessors which can be operated either " o f f line" as stand-alone equipment, or "on line" interfaced directly to the minicomputer. The A10, when originally purchased, was fitted with the EK8 for automatic coordinate registration to provide punched paper tape for aerial triangulation computations. To date the A10 has not been u p g r a d e d ; p a r t l y because the EK8, although an unintelligent device, is capable of digitizing " o n the fly" and thus can be used to obtain limited digital data. The A10 and one of the A8s are located many hundreds of miles away from the central processor. Both instruments are fitted with ahigh-speed paper tape punch which provides the means of transporting high-volume digital data over long distances. Although it is feasible to interface these instruments to the central processor through a telephone land line, this is considered to be uneconomic at the present time. Two of the instruments, one A8 (Fig. 1) and the B8, together with the graphic digitizer table (Fig. 2), which are located with, and interfaced directly to, the central processor via their microprocessors, are also equipped with
Fig. 1. Analogue Stereoptotter with 24 K Microprocessor.
Fig. 2. Graphic digitizer with microprocessor.
207
video screens which have a switch capability enabling them to be used as either alpha/numeric terminals or as graphic screens, using a 256 square dot matrix. These microprocessors are programmed to sample through the u p / d o w n counters, the square wave impulses from the shaft encoders, or in the case of the graphic digitizer table the inbuilt microprocessor, on a continuous basis, so long as the foot switch of the equipment is depressed. In theory t h e y are capable of sampling at the rate of 2400 characters per second, but in practice this is reduced by time filtering to approximately 20--25 per second. The registrations are retained in the machine system, that is a local system of coordinates, and are c o m p u t e d and displayed on the video screen when used as a graphic screen. As each planimetric feature or complete c o n t o u r is digitized, it is displayed and the operator has the ability to either reject or save the data according to his assessment of its correctness. If he chooses to save the data, it is stored on a permanent file on one of the diskettes. On completion of the digitizing of model or plan, the data is transferred to the minicomputer. The microprocessors can be operated in either of two modes, one mode is " t r a n s p a r e n t " where the microprocessor behaves essentially as a non-intelligent Visual Display Unit, the other mode is intelligent where the microprocessor responds to and transmits data to the minicomputer. In its " t r a n s p a r e n t " mode the microprocessor (Fig. 3) is "signed o n " to the minicomputer, enabling it to be identified. In this mode the microprocessor is under the control of the minicomputer, which establishes and allocates file space.
Fig. 3. Microprocessor Videoscreen and Disks. Fig. 4 . 4 8 K Mini Computer.
208 Once the microprocessor has been "signed o n " it reverts to its second mode, that of an intelligent unit, acting together with the minicomputer, and begins to transmit the stored data. The data is transformed b y the minicomputer into the terrain system of coordinates, a mathematical filtering of the data, based on a simple test of the deflection angle at the third point in any trio, is applied, and the data is stored on the 10 megabyte disc.
Processing The processing equipment consists of a Data General Nova 840 (Fig. 4), which was enhanced with various add-on units to enable the storage of highvolume digital data. The c o m p u t e r is m e m o r y - m a p p e d and can be operated foreground/background in time-shared BASIC and F O R T R A N dedicated to a single application. It is disc based and supports in-house "time-sharing" using BASIC thus the full core m e m o r y available in BASIC is available to each user. There are no hard and fast rules applying to operation of the system, b u t usually it is operated in time-shared BASIC during the day, with batch F O R T R A N run overnight. During the daytime, processing the data-acquisition stage is given priority over other users when data is to be transferred from the microprocessors. The system is used for all computations, including commercial processing, j o b costing, payroll, management reports etc., aerial triangulation computations, volume capacity and earthwork calculations. The minicomputer also provides the drive to the flatbed plotter through use of the spooling device. The plot instructions are c o m p u t e d in Fortran overnight, the data having been set up in BASIC during the previous daytime processing. Usually the plotter is left on overnight and commences plotting during the night until it needs operator intervention. The remainder of the plot instructions are left in the spool and provide the data to keep the plotter going during the next day's processing.
Editing Editing of the digital data before any hard-copy plotting is attempted is in all probability the most vital stage in a digital system. The equipment utilised here consists of a Tektronix 4010 storage display screen (Fig. 5) connected to a Nova 4, 32K minicomputer which provides the computing power. Data to be displayed and edited is stored on a 10 megabyte disc which is interchangeable with the 10 megabyte disc on the central processor. The digital data is displayed on the screen and edited by use of the cross hairs. Software has been developed to enable the data to be windowed, deleted, moved, added to, displayed at different scales, automatically squared, contours to be joined and so on. The procedure adopted has been to digitize data either on the stereoplotter or the graphic digitizer table and display it on the video screen, using the
209
Fig. 5. Storage Display Screen.
Fig. 6. Hybrid Flatbed Plotter.
raster scan d o t matrix. The data is transferred to the minicomputer, transformed into terrain units and re-displayed on the storage screen where final editing is completed prior to hard-copy plotting. Presentation
At the present time, flatbed plotting is achieved by using a hybrid plotter (Fig. 6) consisting of an analogue instrument plotting table which has been modified by the addition of stepping motors to the x and y drive spindles (Fig. 7) and interfaced to the minicomputer. The equipment does not have the quality of line work produced by a recognised flatbed plotter; however, this is often unimportant in civil engineer° ing work provided it is capable of maintaining required accuracies. Final cosmetics are still obtained in the conventional way. Software routines have been developed for a wide range of graphic output, from edited and reduced detail or c o n t o u r plots to capacity curves, c o n t o u r generation from random input, gridded control sheets, point plots of field generated "string" data, long sections, cross sections and so on. Ultimately it is intended to install a high-speed flatbed plotter which will enable plans to be produced with a higher presentation quality, so that with the minimum of additional work they can be reproduced and delivered to the client.
210
Fig. 7. Flatbed Plotter Control Interface.
Cost effectiveness Most if not all, private mapping concerns are faced with a restricted capital budget and our c o m p a n y was no exception. The hardware capital cost of the digital system, excluding the original cost of the stereoplotters, has been $A160,000 which is comparable with the price quoted in Australia for one analytical plotter, and this a m o u n t has enabled four analogue stereoplotters to be integrated into the system. Before purchasing equipment, cost effectiveness has to be carefully scrutinised. To be cost effective in the commercial world all equipment purchased must be utilised to the m a x i m u m and must improve upon the efficiency or accuracy of a current procedure. It must also produce a marketable product which is otherwise unobtainable, or a product at least comparable with and at lower cost than existing products (Cleaves, 1978). It is very difficult to assess the cost effectiveness of a digital system. Contrary to many opinions, digital data is not obtained almost as a by-product of the plotting procedure. In fact, the cost of producing digital data can be almost 100% on the cost of producing a graphical result. The true cost effectiveness or advantages are not experienced at the acquisition stage but rather at the user stage. We have found that stereoplotting time for man-made features has increased by up to 25% and that plotting of contours and drainage increases plotting time by 15%. Onto these increased times must be added the time taken to edit the data, which we have found to be around 40% to 50% of the plotting time, including the time needed to edge match model joins and sheet edges. Besides the increase in time necessary to produce a digital map, there is the additional capital investment in the equipment necessary to produce this data, which has to be offset against the cost of the product. However, if the p r o d u c t being produced is a new product, providing data that will be analysed m a n y times by the introduction of new parameters,
211
then it can certainly be considered to be cost effective, particularly if the new parameter permutations are innumerable. There are some cost savings to be made by a mapping organisation with a digital system. These savings are mostly seen in the areas of aerial triangulation and the c o m p u t a t i o n of volumes where the digital system enables rapid on-the-spot checks to be made on data which is otherwise hard to check, or which, when checked, involves costly re-observations. PERFORMANCE, APPLICATIONS AND S T A F F SATISFACTION
As with any new technology it is necessary to evaluate performance against expectation. The expectations for the digital system were threefold: (1) aerial triangulation; (2) engineering c o m p u t a t i o n s ; a n d (3) digital map production. Prior to establishing the hardware necessary to interface analogue instruments to a computer, considerable expertise had been built up in aerial triangulation and engineering computations using batch processing and timesharing bureaux. The transition from external computers to an "in-house" minicomputer had already been made and much of the expertise built up with outside computers was transferred and enhanced with the company's own computer.
Aerial triangulation Computation procedures for strip and two-dimensional block adjustments were firmly established. After purchase of the minicomputer a three-dimensional block adjustment program was developed, based on the xy program developed at Melbourne University by Bervoets (1971). Using a microprocessor " o n line" to the computer, the assembly of models into strips takes place as each model is read, with residuals on the pass points displayed on the video screen. Operator job satisfaction is enhanced and job costs reduced as less mistakes and re-observing of models have occurred. Procedures have now been developed which retain the digital data in the system from the time of original observation until gridded control sheets are returned to the instruments for the scaling of individual models. The'operators are supplied with elements of orientation, together with a model diagram which displays all points to be found on the model in their approximate position, plus the residuals of that model with respect to the final mean values. Flatbed-plotted control sheets for all plans on the job are provided, each sheet having a punch register position plotted which enables the sheets to be accurately joined on a stud register to all the surrounding sheets, thus removing all the problems of edge matching of sheets both on the plotting table and in the drawing office. If required, punch registered overlays for
212
all control sheets can also be provided, with the c o m p u t e r accurately controlling the punch register positions. J o b satisfaction has been considerably enhanced for almost of personnel as a result, with ready acceptance of the benefits that the system provides. Cost savings have also been measured, both in the reduction of personnel directly involved plus an increase of throughput as a result of the increased accuracy and reliability of the products produced during this phase of plan production.
Engineering computations For some years prior to the installation of the digital system, programs had been developed to c o m p u t e volumes, areas, capacity diagrams and earthwork quantities. The data preparation for these types of calculations had been either b y hand or b y point m o d e registration units onto punched paper tape or punched cards, with processing carried o u t on c o m p u t e r time-sharing bureaux. With the installation of the microprocessors " o n line" to the computer, the observation and recording of input data was greatly simplified and checking of the complex shapes is overcome b y displaying the profiles on the video screen at the acquisition stage. In simple cases where one of the surfaces is a plane, such as a reservoir capacity, it is also possible to c o m p u t e volumes and areas as the observation of the data is proceeding, such that there is no time delay between completion of observations and achieving a final result. Hard c o p y of these results is easily obtained b y storing progressive figures in a file in the c o m p u t e r and printing the results on the line printer. The methods of digitizing, joining of two irregular surfaces and computation of volume has been adequately described by Leatherdale and Keir (1979). However, for a number of reasons clients still require their results in terms of horizontal profiles, as represented b y contours, rather than vertical profiles which clearly define the complex shapes found in quarries or open-cut mining. A method of translating data digitized in vertical profiles into horizontal profiles has been developed, which enables these complex shapes to be digitized in the most accurate and economical way and yet satisfy client requirements for area and volumes to be presented in a form most usable by them. Over the past few years all of the major reservoirs which have either been investigated or built in the vicinity of Melbourne, a city of a b o u t three million people, have been surveyed by use of digital photogrammetry. Some five major reservoirs have been, or are being built. Initial capacity tables and capacity diagrams were prepared during the investigation stage, earthwork volumes were c o m p u t e d for the dam wall construction, b o t h the material removed in excavation to bedrock and the materials placed in the wall construction. On completion, as-built plans and final capacity tables and capacity curve diagrams have been prepared.
213
A particular problem on which the digital system is used has been the monitoring of quarries, both during their excavation and afterwards, when they are generally taken over by local councils, who use them as rubbish dumps until refilled. They are then landscaped and returned to the comm u n i t y as parkland or recreational areas. Although an apparently mundane problem, considerable sums of m o n e y are involved as well as safety and public health risks where clay or sand pits are close to, or now within, the suburban sprawl of a major city. The digital system is a fast, economical way of monitoring these problems and enables accurate volume and area figures to be produced which engineers use to specify rate of fill, final desired surface as well as reserves, in terms of remaining space and so forth.
Digital map production Digital map production is the last of the three objectives of the digital system and could n o t be attempted until a successful interface between stereoplotting instruments and computer had been achieved. Development of the procedure, both hardware components and the considerable software requirements, has dominated the activities of the computer room staff. The digital map production procedure has been developed around a modified feature code based on a recommended standard of feature codes developed by the National Mapping Council of Australia (Standards Association of Australia, 1980}. The modified feature code of four digits can be computerextended to the eight-digit feature code required by the standard. The microprocessors have been programmed to provide a p r o m p t facility in the form of a menu display, as an aid to the stereoplotter operators in selecting the feature codes to be used. The stereoscopic models can be digitized in
a
1'500
Fig. 8. Examples of digital mapping.
b
1"2500
214 any sequence and it is not necessary to maintain a particular feature code once selected, although for convenience this is usually attempted. The microprocessors are wired into the footswitch of the instruments so that they are activated as soon as the footswitch is depressed. The m e t h o d of operation is for the stereoplotter operator to key in the feature code and proceed to plot. The data recorded by the microprocessor consists of a pen command activated by the action of the footswitch and the XYZ coordinates of the instrument axis. In this way the data recorded is minimised, with sufficient information to enable identification by the c o m p u t e r at presentation stage. Data is separated into its respective feature code and colour code when it is transferred to the minicomputer. This facilitates editing and flatbed plotting, a~s line sizes and types, symbols etc. are matched up with particular feature or colour codes. The digital system has been used for mapping in the scale range 1:500 to 1:100,000 with magnetic tape supplied to clients for further use. An example of digital mapping is shown at Fig. 8 with 1:500 and 1:2,500 maps produced from the same data. The fundamental aim of this organisation has been to upgrade existing analogue instruments by the addition of microprocessors and related peripherals to acquire digital map data in the most economical way. The minimum data necessary to describe any line is captured by sampling at a high rate, and mathematically reducing this data such that the plot produced from the reduced digital data would faithfully reproduce a plot produced from the table of a stereoplotting instrument. CONCLUSION In development of the digital system our aim has been one of staff involvement in new technology to produce a new product efficiently. Staff must obtain job satisfaction and must not feel their security is threatened by the instrumentation of new technology. It is very difficult to obtain cooperation and feedback from highly skilled stereoplotter operators by telling them that the new equipment can be operated by semi-skilled people. Equally, job satisfaction deteriorates rapidly if all so-called tedious tasks are removed only to be replaced by a lesser number of even more tedious tasks. There is no job satisfaction whatsoever to a highly skilled stereoplotter operator who is asked to spot height a stereo model on a tight grid mesh for the c o m p u t e r to produce contours. He would much rather produce the contours and acquire digits during the process. The fact that this does not lend itself to easy computer processing is of no interest, for computers are supposed to enhance his lifestyle. The social consequences of the new technology have to be considered if retrenchment of skilled staff, or lack of o p p o r t u n i t y to y o u n g eager graduates
215
of academic colleges is to be avoided. It is very difficult for a small c o m p a n y faced with an explosion in costs, particularly labour costs, to consider this last point, b u t if a successful transition to a digital system is to be attempted it has to be considered. After all, a digital system is complex and requires highly intelligent people to develop and operate it. POSTSCRIPT
This paper was originally written in January 1980 and published at the I.S.P. congress in Hamburg in July 1980. Since then, continuous research and development has naturally produced some changes to the system described. In addition, because of the development in microprocessor technology the author can see a major philosophical change in future digital mapping systems. The A.A.M. digital system is what is known as a distributed processing system. The main advantage of such a system is that it is n o t dependent on a single source of intelligence (i.e. the central processor), b u t has local intelligence via the microprocessors at the stereoplotters. Thus the entire system is not brought to a halt in the event of a hardware failure in the central processor. The initial development of this system called for 24K microprocessors, programable in the BASIC language, which were capable of acquiring digital data and transferring it to the central processor. The " o n line" processing p o w e r was limited to displaying the data on the videoscreen in the local coordinate system and to some simple edit routines. These microprocessors have n o w been upgraded to 64K bytes, which are programmable in F O R T R A N and which have a far greater " o n line" processing capability. As such they are n o w able to carry o u t processing tasks previously done by the central processor. Transformations from the local coordinate system of the stereoplotter to the terrain coordinate system are n o w carried o u t in real time at the acquisition stage, together with the real time display of the graphics on the videoscreen. In addition calculations for perspective centre coordinates, model join residuals, areas and volumes, and absolute and relative orientation are all carried o u t in real time b y the microprocessors at the stereoplotters. Thus if the central processor fails at any time due to a hardware fault the acquistion stations (stereoplotters or graphic digitizer tables) can continue to work, storing data on diskettes if required. The major philosophical change as seen b y this author is the disappearance of the central processor altogether from future digital mapping systems. Current microprocessor technology is such that all of the tasks carried o u t by the central processor can be distributed to a series of microprocessors interlinked to enable data to be transferred around the system as and when necessary. Thus the edit stations can be microprocessor supported, and the flatbed or automatic plotting table can be microprocessor controlled. A b o u t the only c o m p u t a t i o n which couldn't be handled b y a microprocessor at present would be a Bundle block adjustment.
216 The major advantage t o such a distributed system will be t h a t hardware failure to any part of the system will only affect t h a t part, with all o t h e r functions o f th e system operating normally. As microprocessors are relatively inexpensive it will p r o b a b l y be possible t o carry a c o m p l e t e spare system which can be plugged in t o any part of the system in the event of a hardware failure. REFERENCES Leatherdale, J.D. and Keir, K.M., 1979. Digital methods of map production. Photogramm. Rec., 9 (54): 757--772. Cleaves, R.P., 1978. Mini-Map (A semi-automated mapping system). 21st Austr. Surv. Congr., Adelaide, April, 1978, pp. 135--142. Bervoets, S.G., 1971. Horizontal block adjustment, Austr. Sure., 23 (5): 304--317. Standard for the exchange of feature coded digital mapping and charting data on Magnetic Tape. DR 79178 Standards Association of Australia, 1980. -