- Email: [email protected]
Detection of latent bloodstains at fire scenes using reflected infrared photography
Accepted Manuscript Title: Detection of Latent Bloodstains at Fire Scenes Using Reflected Infrared Photography Author: Belinda Bastide Glenn Porter Adrian Renshaw PII: DOI: Reference:
S0379-0738(18)30660-1 https://doi.org/doi:10.1016/j.forsciint.2019.109874 FSI 109874
To appear in:
FSI
Received date: Revised date: Accepted date:
28 August 2018 29 June 2019 23 July 2019
Please cite this article as: B. Bastide, G. Porter, A. Renshaw, Detection of Latent Bloodstains at Fire Scenes Using Reflected Infrared Photography, Forensic Science International (2019), https://doi.org/10.1016/j.forsciint.2019.109874 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.
Detection of Latent Bloodstains at Fire Scenes Using Reflected Infrared Photography
Western Sydney University, School of Science & Health, Richmond, Australia.
us
a
cr
ip t
Belinda Bastidea*, Glenn Porterb, Adrian Renshawc
University of New England, School of Humanities, Arts & Social Sciences, Armidale, Australia.
c
Western Sydney University, Science of Science & Health, Richmond, Australia.
Ac ce pt e
d
M
an
b
Declaration of Interest: The are no interests to declare for this research.
Original Research Article 1
Page 1 of 31
Abstract Bloodstain evidence is an element of crime scene investigation often found at
ip t
scenes involving violence. Setting fire to the scene is a method sometimes used by offenders of crime in an attempt to conceal evidence. Fire often produces thick soot as a
cr
by-product of the combustion and has the potential to cover bloodstain patterns
us
rendering them latent. There is limited published material offering a method of detecting bloodstains hidden beneath dense soot deposits caused by fire. This project employed a
an
modified digital single-lens reflex (SLR) camera to investigate the application of reflected infrared photography to detect latent bloodstain evidence beneath varying deposited
M
overlaying soot densities. The potential of this technique was examined by photographing blood samples beneath soot from a scaled fire simulation. A qualitative
te
d
evaluation was completed by comparing images taken of a series of samples using both reflected infrared and standard visible light photography and corroborated with
Ac ce p
quantitative image analysis to support the findings. Results indicate that infrared photography can reveal latent bloodstains beneath a dense layer of soot in excess of ρ2.3 (550 nm) density with substantial clarity. The success of this technique is dependent on specific optical and specimen parameters. These parameters include i) the reflective properties of the background surface, ii) the spectral absorption properties of blood and iii) the ability of infrared wavelengths to transmit through the soot layer. Reflected infrared photography may provide crime scene examiners with a specialised field recording method that is easily executed and non-destructive to assist in visualising and locating latent bloodstain patterns beneath dense layers of soot. 1
Page 2 of 31
Keywords; infrared photography, bloodstains, latent evidence, forensic photography,
Ac ce p
te
d
M
an
us
cr
ip t
crime scene, blood evidence.
2
Page 3 of 31
Introduction Blood and bloodstain evidence recovered from scenes of crime involving violence is regularly used in forensic investigation for crime scene reconstruction and forensic
ip t
identification purposes [1-5]. A range of forensic science methods are employed to detect the presence of blood at sites involving violent crime. However, when bloodstain evidence is
cr
concealed due to the event of fire and covered by a deposit of acrid soot, detection using
us
more traditional trace blood detection techniques may not be effective [4].
Forensic aware offenders may employ methods of evidence concealment designed
an
to leave no trace evidence of the crime for forensic investigators to discover. Arson, or deliberate setting fire to property, can be a modus operandi to conceal evidence resulting
M
from violent crimes including homicide and has the potential to destroy physical and biological trace evidence. However, construction fires are often extinguished by fire
te
d
brigades before the structure and any trace evidence is entirely destroyed. Arson investigation is a multidisciplinary field requiring knowledge from a diverse
Ac ce p
range of different studies, including; i) chemistry (properties of accelerants and combustible material), ii) fire science (dynamics of fire), iii) crime scene investigation and iv) firefighting practices. This nexus of considerably different knowledge sources creates a high degree of complexity for forensic investigation that have involved fire. Smoke and sludge from a fire can cover and preserve bloodstain evidence on walls,
ceilings and other surfaces by establishing a protective layer [5]. There are a variety of optical and chemical techniques currently used within regular crime scene investigation and forensic photography to enhance bloodstain evidence at more conventional crime scenes [6] that usually do not involve fire. Chemical blood reagent techniques including; luminol, 3
Page 4 of 31
rhodamine 6G, amido black, leuco crystal violet, Hungarian red and various optical enhancement techniques are also employed [7]. However, excessive soot overlaying bloodstains may cause significant interference when chemical reagents are used in the
are often ineffective due to the carbon layer over the bloodstains.
ip t
recovery of this form of evidence [9] and non-destructive optical enhancement methods [6]
cr
Prior to the previous decade, any physical and biological remnants covered in soot
us
from fire were regarded as ineffectual [9-10]. It was often believed by crime scene examiners that the recovery of such evidence is not possible or too difficult. Consequently,
an
this situation has caused investigators to overlook the recovery of bloodstain evidence remaining in fire scenes [11]. This thinking was based on a suggestion that evidence exposed
M
to fire conditions would degrade and not permit more general investigation techniques to be effective or would not have any meaningful outcomes [9,11]. More recent studies
te
d
however, suggest various forensic techniques may be employed to visualise, enhance and collect fingermarks, DNA and blood evidence from fire affected sites including when the fire
Ac ce p
involved flashover [11-12]. The successful recovery of bloodstain evidence from fire scenes may lead investigators to better understand the events that occurred prior to the fire through bloodstain pattern analysis (BPA) and offer further identification evidence using DNA techniques from recovered trace blood. Photography is often considered as a primary source of evidence preservation for
several forms of physical evidence due to its non-destructive nature [6,13-14]. Infrared radiation consists of longer wavelengths [15] and has the ability to penetrate through some visually opaque medium resulting in latent material beneath becoming visible [6,16]. Infrared radiation also has the potential to discern between two different objects that 4
Page 5 of 31
appear optically similar under visible light due to the difference in spectral response from the materials irradiated by infrared radiation [16-17]. This metameric effect has significant potential value for forensic investigation at fire scenes.
ip t
When material is photographed using infrared radiation there are three possible recording outcomes; i) the material may absorb infrared radiation and become darker, ii)
cr
the material may reflect infrared radiation and become lighter or, iii) the material may
us
transmit infrared radiation and become transparent [18]. Infrared photography requires infrared sensitive film or digital cameras suitable (or modified) for recording infrared and a
an
light source containing infrared radiation [17-20].
When bloodstains are deposited onto various surfaces at crime scenes, they
M
immediately begin to oxidise or dry on surfaces. When blood dries it undergoes a change of colour from a vibrate red to a dark brown hue due to the chemical change as the oxidisation
te
d
process occurs. Cadd et al [21] explains the oxidisation of the haemoglobin changes composition from haemoglobin (Hb) to oxyhaemoglobin (HbO2) which then changes to
Ac ce p
metaemoglobin (metHb) and finally fully denatured into hemichrome (HC) [21]. Horecker [22] cites early experimentation work on the absorption of infrared radiation by haemoglobin and its derivatives by Merkelbach [23] as early as 1935. It was found that oxyhaemoglobin (HbO2) and methaemoglobin (metHb) can absorb a broad range of wavelengths, while carboxyhaemoglobin is unable to absorb infrared radiation [22]. Red blood cells are tightly packed with haemoglobin, consisting of approximately
one million haemoglobin molecules in one erythrocyte [24]. MetHb formation after heat exposure indicates the decomposition of blood components throughout an ageing process [7]. The exposure of blood to fire conditions can speed up this decomposition and ageing 5
Page 6 of 31
process, causing HbO2 material to be converted to metHb at an increased rate. This rate of decomposition is dependent on the exposure temperatures within the fire. Kuenstner & Norris [25] discovered the peak absorption for metHb and HbO2 occurs at approximately
ip t
630 nm and 930 nm respectively, with metHb absorbing approximately 10-times more infrared radiation than HbO2 [25]. The production of metHb from heat exposure and its
cr
greater infrared absorption properties may have a positive effect on the ability to record
us
latent bloodstains beneath soot when using reflected infrared photography.
Modelling the conditions of construction fires can provide an insight into the types of
an
evidence recovery techniques used by investigators within a forensic science context. The objective of conducting a scaled model experiment is not to replicate a full fire scenario,
M
rather it is to create easily controlled experimental fire conditions to determine the recoverability of bloodstain evidence using reflected infrared photography. The scaled
te
d
experiment will allow control parameters to be actively regulated including i) exposure temperature, ii) fuel load, iii) ventilation properties, iv) bloodstain location and v)
Ac ce p
determination of soot deposit density. The use of a quasi-steady fire model will profile accurate scaling of the burning rate and room environment produced by fire [26-27]. This model will provide replication of gas and wall temperature, gas species concentration and radiation flux at homologous locations when correct scaling laws are applied [27-28].
Materials & Methods
The experimental design consisted of several stages including; i) the fire modelling, ii) photographic recording and imaging, iii) measuring soot density at various hot layers and at differing wavelengths, iv) the examination of the images and v) analysis of results. 6
Page 7 of 31
Fire Modelling The scaled enclosure model in this project uses the theoretical design of a quasi-
ip t
steady model. This scaling scheme is used to replicate similar geometric and free burn behaviour from one scale to another [27]. The use of a quasi-steady model allows burning
cr
rate and gas temperature rise in realistic enclosure fires to be within those observed in free
us
burn cribs.
The enclosure model was designed from a full-scale bedroom with a single doorway
an
opening. The model was scaled to a 1:5 ratio of a full-scale building enclosure using plywood foundations. The lining of the enclosure was designed using plasterboard wall and ceiling
M
and metal tile flooring. The dimensions for both full scale and scaled enclosure buildings are shown in Table 1. The internal dimensions were measured whilst including the plasterboard
the scaled model.
te
d
wall linings, therefore small adjustments were made to allocate for plasterboard linings in
Ac ce p
This experiment uses a wooden crib fuel source and follows the burning regime using stick thickness. The stick dimensions used in the crib fuel are shown below in Table 2. The wood crib was constructed out of small clear pine wooden dowel, 2.5 cm in diameter and 400 kg/m3 density. Figure 1 shows the scaled modelled fire in action. This experiment used 200µL defibrinated horse blood sourced from a commercial
biological supplier, and was pipetted into specific locations of the gyprock on the rear wall of the scaled model. To place the bloodstains in the same location on the gyprock for each sample, a Perspex stencil was used as described by Farrar et al [20] (Figure 2). The locations of the 30 bloodstains were determined by drilling holes in the Perspex in a grid formation, 7
Page 8 of 31
parallel to the thermocouples places along the height of the enclosure. Each 200µL of blood was deposited onto each modelled fire wall using the grid for consistency of placement and location recovery. Figure 5 denotes the positioning of each bloodstain within different heat
cr
safe ignition of the fire and easy extinguishing if required.
an
us
ip t
layers. The scaled enclosed model was placed in an outdoor area on a flat hard surface for
te
Fuel Crib Ignition
d
M
walls>
To create replicable burning behaviour, the ignition of the fuel crib occurred in the
Ac ce p
same manner in each repetition. This allowed the same amount of heat energy to be released from the fuel. Ignition of the fuel crib was conducted using a kerosene-soaked rag. One piece of 200 mm x 200 mm cotton rag was soaked in 50 ml of kerosene and placed in the bottom of the fuel crib. The rag was then ignited using a hand-held fire starter gun. Once the fuel-soaked rag was ignited it then provided an initial heat flux to cause the wood dowel crib to ignite and become the dominate source of fuel source and heat energy.
8
Page 9 of 31
Fuel Crib Extinction A suffocation technique was used to extinguish the enclosure fire in a safe manner. As the fire was ventilation controlled, a sudden lack of oxygen caused fire to smother and
ip t
extinguish. This was achieved by closing the doorway opening and starving the fire of oxygen. Once the doorway was closed for a significant amount of time, (≈5 minutes), the
cr
door was reopened and a garden hose was used to spray water onto the crib. This small
us
amount of water caused any smouldering wood to become saturated. The hose was set on a low-pressure mist with a diameter small enough to cover the crib and not cause any
an
splashing onto the soot deposited walls within the enclosure.
M
Determining Soot Density
Transmission measurements recorded within this project examined the ability of
te
d
radiation to pass through a medium. Transmittance is the fraction of incident light at a specified wavelength that passes through a sample [29]. Percent transmission is expressed
Ac ce p
in equation form:
=
100
Equation 1: Percentage transmission.
The speed at which radiation wavelengths can pass through a medium depends upon the optical density of the material. The optical density of a material relates to the sluggish tendency of atoms to maintain the absorbed energy of an electromagnetic wave in the form 9
Page 10 of 31
of vibrating electrons before reemitting it as electromagnetic disturbance [29]. The more optically dense a material, the slower a wavelength will pass through it. The optical density
ip t
of a material is expressed as:
cr
= log
us
Equation 2: Optical density.
an
The density of the overlaying soot within the enclosure was determined using an Ocean
M
Optics HR2000 spectrophotometer. Small pieces of rectangular glass (microscope slides) were attached to the walls surface at the same heights where the bloodstains were located,
d
at 2 cm, 10 cm, 20 cm, 30 cm, 40 cm and 50 cm. Transmission measurements from these
te
slides at each height location were collected 5 times each and the mean transmission value determined. Transmission of each height was calculated using Equation 1 [29], then these
Ac ce p
measurements were converted into optical density measurements using Equation 2 [29]. The spectral transmission of the soot indicated transmission of visible light through to infrared. The longer infrared wavelengths were able to penetrate through the soot aggregate more efficiently than other wavelengths in the visible spectrum (see Figure 3) when optical densities were measured. A lower optical density result in the infrared wavelengths (900 nm) and in effect allows greater transmission in each of the fire layers (heights in the modelled fire). While the physical deposit of soot increases as the samples are placed in the higher heat layers of the fire, the optical density of soot also varies with
10
Page 11 of 31
different wavelength. Figure 3 also demonstrates the relationship between the relative density of the overlaying soot and wavelength.
ip t
us
cr
the scale enclosure.>
Camera Modifications and Spectral Sensitivity
an
The project used a modified Canon digital single-lens reflex (SLR) camera that was converted to record exclusively within the infrared region (>820 nm). The camera was
M
modified by removing the infrared blocking filter over the sensor and replacing it with an infrared transmission filter (equivalent to a Wratten 87C filter).
This longpass filter
te
d
transmits longer wavelengths above 820 nm while rejecting shorter wavelengths less than 820 nm. Figure 4 provides the spectral transmission of the optical filter fitted over the
Ac ce p
camera sensor.
Standardising Exposure Values and Camera Settings
11
Page 12 of 31
To enable a meaningful comparison between the visible light and infrared images, the images must be captured using exposures of equivalent exposure density or pixel brightness values. The difficulty when comparing images taken with different radiation
ip t
sources and camera spectral sensitivity is determining equivalent camera exposures for each radiation source that result in the same image density range. In other words, taking
cr
two different camera exposures using different radiations sources, different camera spectral
us
sensitivity and differing optical filtration, to achieve exposures that can be considered as alike. A calibration factor between the differing overall sensitivity of each sensor was
an
determined.
The ‘equivalent’ camera exposures were tested using a Kodak Q13 ‘Grey Scale’
M
reference standard with each camera (visible light and infrared) and using the same tungsten lighting with even illumination. Tungsten lighting was selected because it has a
te
d
spectral distribution characteristic that includes both visible light and infrared radiation. Testing exposure values consisted of taking a series of photographs of the subject with an
Ac ce p
attached Kodak Q13 ‘Grey Scale’ reference standard with each camera using a range of aperture settings.
The eyedropper tool in Adobe PhotoshopTM was used to measure the pixel
brightness of the ‘M’ segment of the Kodak Q13 greyscale reference. The ‘M’ segment has an 18% reflectance with an optical density value of ρ0.7. The eyedropper tool was used to measure the numerical value of pixel brightness between 0 and 255 for 8-bit images. The exposure values for each photographic method were chosen by measuring the ‘M’ segment and choosing the settings that gave the closest value to 18% grey at a 128-brightness value [20]. Table 3 provides the results of the camera exposure calculations for each type of 12
Page 13 of 31
photography. All photography throughout the experimental work was taken at this novel standardised exposure method and the lighting remained consistent during the calibration
ip t
test.
cr
us
Visual Examination and Image Analysis
Visual examination of the test photographs was conducted on a calibrated computer
an
monitor. The visual examination purpose was to determine the soot density parameters where the latent bloodstains became visible. The visual examination was completed for all
M
photographs taken using both reflected infrared and standard visible light photography. Image analysis used a combination of Adobe PhotoshopTM and ImageJ TM software to
d
quantify the effectiveness of infrared photography as a technique to visualise latent
te
bloodstains beneath the soot layers. PhotoshopTM was used to prepare the photographs for
Ac ce p
analysis and ImageJ was used to generate pixel brightness values from each image. Data from the image analysis was compiled on all experimental and control photographs using both the standard and infrared cameras. Each photograph was examined in Photoshop and using black tape outlining the
sample plasterboard as a guide, each image was cropped and resized to 2500 x 2077 pixels. The images remained in an upright orientation as indicated by the small ‘North’ indicator placed on each sample. The image was then converted into an 8-bit image and all colour information was discarded to maintain pixel brightness consistency throughout the analysis. This was completed by selecting Image > Mode and by selecting both ‘Greyscale’ and ‘8 13
Page 14 of 31
Bits/channel’. The image was then saved by using the File > Save As function. All images were saved in JPEG format and were set to maximum quality. Each image was opened in ImageJ and the line tool was selected to draw a linear line
ip t
in a horizontal orientation across the pixels where the blood drops were located to ensure data from all pixels across the width of the image was generated. By selecting Analyse > Plot
cr
Profile, a plot profile line was graphed by the software and the pixel brightness values were
us
saved into a spreadsheet. Further readings were taken for each sample repeat by drawing vertical lines throughout the image where the blood droplets were located. This generated
an
11 individual pixel brightness readings where the bloodstains were located. The plot profile is generated from the pixel brightness values contained within the image. Each pixel is
M
assigned a grey value between 0 and 255 representing the grey scale of each individual
Experimental Controls
te
d
pixel.
Ac ce p
Control measurements were taken of the pixel brightness values of the background white paint and the bloodstains to compare the spectral difference between the two objects and determine if there was significant difference between the two. This was conducted using both visible and infrared images captured of each of the samples. The images were examined in Adobe PhotoshopTM using the eyedropper tool to determine the pixel brightness of both the background and the blood. These measurements were then used in the analysis including an ANOVA evaluation to statistically represent any variance between the visible and infrared images [8]. Despite using very specific burn modelling controls including replica fuel loads and ignition processes, fires may still produce variations in burn 14
Page 15 of 31
patterns and soot deposition. This project used four model fires with five samples in each burn layer resulting in 20 samples per burn. The statistical modelling used a mean value for
ip t
all model fire replications to counter burn pattern variability.
Results
cr
Experimental samples were photographed using reflected infrared and standard
us
visible light photography techniques using the calibrated exposure with consistent illumination and relative exposure values. Several elements were examined to determine
an
whether infrared was an effective forensic imaging method to visualise bloodstains beneath dense soot deposits resulting from a fire. The results were evaluated using the following
i)
M
information and data;
A qualitative assessment using a comparison of images viewed on a
ii)
te
d
calibrated monitor,
Consideration of the relative density of the soot on each heat layer
Ac ce p
(expressed as height in the scale-model burn), iii)
A quantitative assessment of the difference between the pixel brightness readings between the soot and blood+soot deposits at each heat layer or height (also considered as contrast between the two areas).
The four replicants produced consistent results for burns 1,2 and 4. However burn 3
resulted in significantly less soot deposit due to the unpredictable nature of fire. Nevertheless, the other three burns provided consistent results to draw conclusions while the third burn may be considered as an outliner. The only burn height that was of interest to
15
Page 16 of 31
examine whether infrared was more effective than standard visible light was the 50 cm scale-model height due to the greater deposit of soot in the hotter layer. The examination of the relative soot densities at different wavelengths indicated
ip t
there is a significant difference for the ability of differing wavelengths to penetrate through soot deposits. Figure 3 indicates several elements to support this outcome; i) soot density
cr
increases within the hotter (higher) fire layers and, ii) longer wavelengths lower the relative
us
density which results in greater potential penetration capacity. Figure 3 reveals a relative optical density at 50 cm height of ρ2.9 for 400 nm wavelength, ρ2.3 at 550 nm, ρ1.8 at 700
an
nm and ρ1.3 at 900 nm. This result suggests infrared radiation has an improved potential to image through soot deposit layers than visible light. All other fire layer heights have a similar
M
trend however these heights are less important to this project.
Since relative density is a logarithmic value the difference displayed in Figure 3 are
te
d
significant. At 50 cm height the difference between 400 and 900 nm wavelengths are more than 30-times greater, while between 550 nm approximately 8-times relative density
Ac ce p
difference. This is a significant result when considering the potential of infrared photography’s ability to penetrate the covering soot deposit to record underlying bloodstains.
The visual assessment of standard visual light photography indicated it was difficult
to see the bloodstains with soot deposits >ρ1.5 relative density (550 nm) while infrared photography results show a substantial improvement in the visualisation of bloodstains at increased relative density levels (>ρ2.3 at 550 nm which is approximately 8-times greater). Further quantitative assessment of the visible light and infrared images using pixel brightness values (0-255 range) measured directly from each image also indicated 16
Page 17 of 31
improvement in the visualisation of bloodstains at higher relative densities of overlaying soot. Figure 6 plots the pixel brightness values from each burn at differing heights (fire hot layers). These plots display the difference between the brightness value (using a calibrated
ip t
exposure as indicated in the methods section between each imaging system) of the soot only to the pixel value of the soot+blood. This difference is considered as contrast within the
us
range within silver halide or film photographic technology.
cr
photographic recording capacity and is analogous to Dmax-Dmin expression of contrast
Figure 6 plots, with the exclusion of Burn 3 considered as an outlier, demonstrates
an
no advantage for infrared imaging at the lower heights and density deposit sections. This is to be somewhat expected because infrared photography is generally a lower contrast
M
recording method than standard white light photography, for general usage. This is certainly the case when comparing the contrast difference between the visible light and infrared
te
d
photography at those lower density coverage heights. However, as seen at the 50 cm height (and to some degree 40 cm height too), there is a significant change in contrast towards an
Ac ce p
improvement of contrast for the infrared recording. Greater pixel difference or contrast produces greater visibility. The pixel brightness values indicate a significant improvement in visualisation of blood under soot density >ρ2.3 (550 nm).
17
Page 18 of 31
ip t
different hot layers. Note; burn 3 is an outlier.
Single-Factor Analysis of Variance
cr
The results of ANOVAs performed on the mean pixel brightness data to determine
us
whether there were any significant differences (p<0.05) between the blood and soot values for visible and infrared images are displayed in Tables 4 and 5. These results indicated that
an
significant variation between the blood and soot values occurred for all heights within the enclosure model. However, the calculated F value was much greater than the critical value
M
for the infrared images at 50 cm height. This indicates a greater magnitude of variance between the blood+soot and soot recorded brightness when captured using reflected
te
d
infrared radiation photography. The ANOVA results further supports the previous findings.
Ac ce p
Discussion
Qualitative and quantitative analysis indicated that when soot density increased, the
difference between standard visible light photography and reflected infrared photography was affected. At the greatest soot density in the upper hot layers (50 cm height) of the scaled-model fire, bloodstains photographed with standard visible light photography methods failed to record the bloodstains beneath the soot with clarity. However, bloodstains photographed using reflected infrared photography at soot density >p2.3 (550 18
Page 19 of 31
nm) significantly improved the visualisation of the bloodstains. Soot deposits ≥ρ1.5 were most effective at concealing bloodstains using standard visible light photography. These soot deposits were located in the hot layer of the scaled enclosure (40 - 50 cm height).
ip t
Soot deposits between ρ0 and ρ1.0 density were the least effective at concealing bloodstains using standard visible photography. These deposits were located in the lower
cr
region of the scaled enclosure (0 - 20 cm height). Experimental work involving measuring
us
the relative optical density of overlaying soot indicated longer wavelengths lowered the relative optical density of overlaying soot. This means that longer wavelengths like infrared
an
has greater potential to penetrate through the material and record substances like blood underneath the soot. In similar fashion to Farrar et al [20] description of how blood is
M
visualised underneath painted surfaces, the visualising of blood beneath soot requires the radiation source (infrared) to transmit through the overlaying soot, be absorbed by the
te
d
bloodstain rendering it darker [25] and then reflect off the wall background to produce contrast resulting in improved visualisation. With dense soot deposits >ρ2.5 wavelengths
Ac ce p
within the visual spectrum will have a reduced ability to penetrate through the overlaying soot resulting in ineffectual visualisation of bloodstains and possibly remaining latent. Infrared photography combines both transparent [18] properties of the soot and absorption of the blood [25] making it a more suitable option for dense overlaying soot. Statistical analysis using ANOVA compared mean pixel brightness values of
blood+soot and soot from standard visible light and reflected infrared images. Reflected infrared photography exhibited significant variance between the tonal values for blood and soot for the higher soot densities. These results provided further corroborate the findings
19
Page 20 of 31
that reflected infrared photography has the potential detect latent bloodstains beneath overlaying soot from arson scenes. The scaled enclosure model used in this project was designed to replicate fire
ip t
behaviour and conditions experienced in full-scale scenarios. The engineered design of the model ensured a broad soot density gradient was achieved such that bloodstains could be
cr
concealed. This included using a calculated, easily combustible fuel source and providing
us
adequate ventilation for control of combustion. Furthermore, modelling fire behaviour provides experimental replication and a model is easy to construct with the availability of
an
materials and inexpensive cost. Using a scaled enclosure built with strong foundations to withstand multiple fire simulations allowed the investigation to achieve results that can be
M
effectively applied to full scale fire scenarios.
The results from this work are relevant to the current forensic field as they provide
te
d
crime scene examiners with a method of detecting bloodstains situated beneath soot. Other techniques investigated for this type of evidence recovery include traditional techniques as
Ac ce p
indicated by Tontarski et al., [9]. The use of traditional techniques was successful in detecting bloodstains beneath soot, however interactions between the chemicals and soot caused difficulties. Results obtained within this work indicate that it is also possible to photograph bloodstains beneath dense overlaying soot using non-destructive reflected infrared photography method.
Conclusion Reflected infrared photography was successful in detecting latent bloodstains beneath dense overlaying soot. This success was achieved by combining the spectral 20
Page 21 of 31
responses of four parameters necessary for this technique including; i) the increased infrared absorption properties of blood after exposure to heat, ii) the infrared transmission and penetration of soot layer, iii) the infrared reflection of the painted background and iv)
ip t
the sensitivity of the digital camera and spectral output of the light source. The combination of these individual infrared responses of blood, soot and background, causes a difference in
cr
tonal values or contrast recorded between bloodstains and overlaying soot, hence latent
us
bloodstains may become visible.
The qualitative comparison between images using standard visible light photography
an
and reflected infrared photography revealed the infrared techniques were able to detect bloodstains beneath dense overlaying soot (and penetrate the soot layer). This qualitative
M
result was confirmed by further quantitative and statistical analysis. Image analysis was used to generate pixel brightness profiles of the images and graphically demonstrate the
te
d
ability to visualise the clarity of bloodstains once increasing soot densities were applied. The difference in mean pixel brightness between blood+soot and soot deposits were verified
Ac ce p
using ANOVA statistical analysis that indicated a greater variance and visualisation in images using reflected infrared photography at more dense soot coverage. The overall results indicated standard photography techniques could no longer
visualise bloodstains after soot densities of ρ1.5 (550 nm) were achieved. Reflected infrared photography was able to produce effective clarity and visualisation of bloodstains beneath soot density in excess of ρ2.3 (550 nm). Since density is a logarithmic scale, these differences are significant.
References 21
Page 22 of 31
[1] Cadd S., Li B., Beveridge P., O’Hare W.T., Campbell A., Islam M., (2016) The non-contact detection and identification of blood stained fingerprints using visible wavelength reflectance hyperspectral imaging: Part 1. Sci Justice. 56 181-190.
ip t
http://dx.doi.org/10.1016/j.scijus.2016.01.004 [2] Cadd S., Li B., Beveridge P., O’Hare W.T., Campbell A., Islam M., (2016) The non-contact
us
reflectance hyperspectral imaging: Part 2. Sci Justice. 56 191-200.
cr
detection and identification of blood stained fingerprints using visible wavelength
an
http://dx.doi.org/10.1016/j.scijus.2016.01.005
[3] Barni, F., Lewis, S., Berti, A., Miskelly, G., Lago, G., (2007) Forensic application of the
M
luminol reaction as a presumptive test for latent blood detection. Talanta. 72(3) 896-913. https://doi.org/10.1016/j.talanta.2006.12.045
d
[4] Vandenberg, N. Oorschot, R.A., (2006) The use of Polilight® in the detection of seminal
te
fluid, saliva and bloodstains and comparison with conventional chemical-based screening
Ac ce p
tests. J Forensic Sci. 51(2) 361-70. https://doi.org/10.1111/j.1556-4029.2006.00065.x [5] Redsicker D., O’Connor J., Practical fire and arson investigation. CRC Press, Boca Raton, 1997.
[6] Porter G., Forensic Photography in Rennie J., et al (Eds) McGraw-Hill Yearbook of Science & Technology. McGraw-Hill, New York 2014 pp124-129. [7] James S.H., Kish P.E., Sutton P.T., Principles of bloodstain pattern analysis: theory and practice. CRC Press, Boca Raton, 2005. [8] Sokal R.R., Rohlf F.J., Biometry: The principles and practices of statistics in biological research, 3rd Edition. W. H. Freeman, 1994.
22
Page 23 of 31
[9] Tontarski K.L., Hoskins K.A., Watkins T.G., Brun-Conti L., Michaud A.L., (2009) Chemical enhancement techniques of bloodstain patterns and DNA recovery after fire exposure. J Forensic Sci. 54(1) 37-48. https://doi.org/10.1111/j.1556-4029.2008.00904.x
ip t
[10] Bradshaw G., Bleay S., Deans J., NicDead N., (2008) Recovery of fingerprints from arson scenes: Part 1 - latent fingerprints, J Forensic Ident. 58(1) 54-82. ISSN: 0895173X.
cr
[11] Deans J., (2006) Recovery of fingerprints from fire scenes and associated evidence. Sci
us
Justice. 46(3) 153-168. https://doi.org/10.1016/S1355-0306(06)71589-1.
[12] Moore J., Bleay S., Deans J., NicDead N., (2008) Recovery of fingerprints from arson
an
scenes: Part 2 - fingerprints in blood. J Forensic Ident. 58(1) 83-108. ISSN 0895173X. [13] Porter G., (2011) A new theoretical framework regarding the application and reliability
M
of photographic evidence. IJE&P, 15(1) 26-61. https://doi.org/10.1350/ijep.2011.15.1.367. [14] Porter G., (2007) Visual culture in forensic science. Aust J For Sci. 39(2) 81-91.
te
d
https://doi.org/10.1080/00450610701650054 [15] Robinson, E., Crime scene photography. Academic Press, Burlington 2007.
Ac ce p
[16] McKechnie M.L., Porter G., Langlois N., (2008) The detection of latent residue tattoo ink pigments in skin using invisible radiation photography. Aust J For Sci. 40(1) 65-72. https://doi.org/10.1080/00450610802047580 [17] Denny K.A., (2015) Recognising changes in visual representation of clothing in CCTV imaging. J Criminological Research, Policy & Practice. 1(4) 233-238. https://doi.org/10.1108/JCRPP-08-2015-0033 [18] Porter, G., Photography & optical enhancement of physical evidence. in Jamieson A. & Moenssens A. (eds.), Wiley encyclopedia of forensic science. 2nd edition. Wiley & Sons. New York. 2013. 23
Page 24 of 31
[19] Verhoeven G., (2008) Imaging the invisible using modified digital still cameras for straightforward and low-cost archaeological near-infrared photography. J Archaeol Sci. 35(12) 3087-3100. https://doi.org/10.1016/j.jas.2008.06.012.
ip t
[20] Farrar A., Porter G., Renshaw R., (2012) Detection of Latent Bloodstain Patterns Beneath Paint Using Reflected Infrared Photography, J Forensic Sci. 57(5) 1190-1198.
cr
https://doi.org/10.1111/j.1556-4029.2012.02231.x
us
[21] Cadd S., Li B., Beveridge P., O’Hare W.T., Campbell A., Islam M., (2018) Age
determination of blood stained fingerprints using visible wavelength reflectance
an
hyperspectral imaging. J.Imaging. 4 (12) 141 1-11. doi:10.3390/jimaging4120141 [22] Horecker B. L., (1943) The absorption spectra of haemoglobin and its derivatives in the
M
visible and near infrared regions. J Biol Chem. 148 173-183.
[23] Merkelbach O., (1935) Infrared absorption and infrared photography of normal and
te
d
carbon monoxide (luminous gas) poisoned blood. Schweiz Med Wschr, 65 1142. [24] Botonjic-Sehic E., Brown C.W., Lamontange M., Tsaparikos M., (2009) Forensic
Ac ce p
application of near-infrared spectroscopy: Aging in bloodstains. Spectroscopy. 24(2) 42,4448. ISSN: 08876703
[25] Kuenstner J.T., Norris, K. H. (1994) Spectrophotometry of human hemoglobin in the near infrared region from 1000 to 2500 nm. J Near Infrared Spec. 2(2) 59-65. https://doi.org/10.1255/jnirs.32
[26] Almirall J., Furton K.G., Analysis and interpretation of fire scene evidence. CRC Press, Boco Raton. 2004. [27] Croce P.A., Xin Y. (2005) Scale modeling of quasi-steady wood crib fire in enclosures. Fire Safety J. 40(3) 245-266. https://doi.org/10.1016/j.firesaf.2004.12.002 24
Page 25 of 31
[28] Jolly S., Saito K. (1992) Scale modelling of fires with emphasis on room flashover phenomenon. Fire Safety J. 18(2) 139-182. https://doi.org/10.1016/0379-7112(92)90036-C [29] Jacobson R.E., Ray S.F., Attridge G.G., Axford N.R., The manual of photography. 9th
Ac ce p
te
d
M
an
us
cr
ip t
Edition. Focal Press, Oxford. 2000.
25
Page 26 of 31
1:5
1:5 + Plasterboard*
Length of Room (mm)
3130
626
646
Width of Room (mm)
3230
646
2560
512
830
166
2030
406
666
cr
532
us
166 406
an
Height of Room (mm) Width of Doorway(mm) Height of Doorway(mm)
ip t
Full Scale
Ac ce p
te
d
M
Table 1 - Dimensions of full scale and scaled enclosure model (*measurements for outside perimeter + 20 mm width of plasterboard lining of the model).
26
Page 27 of 31
Dimensions 150 mm
Width þ
25 mm
Layers N
10
Sticks per layer n
3
cr
ip t
Length l
Ac ce p
te
d
M
an
us
Table 2 - Stick dimensions used in crib fuel source.
27
Page 28 of 31
Lens
50 mm Macro
50 mm Macro
ƒ/stop
8.0
8.0
Shutter speed
1/20th sec
1/125th sec
ISO
100
100
White Balance
Custom
Custom
U distance
2.0 m
2.0 m
EV (Exposure Value)
10.3
13
Light source
Tungsten
ip t
Visible Camera
us
cr
IR Camera
an
Tungsten
Ac ce p
te
d
M
Table 3 – Calibrated camera settings used throughout the experiment
28
Page 29 of 31
Sample Height
n
Mean Blood Value
Mean Soot Value
F
P-Value
F crit
50 cm
20
20.3
51.85
13.41
0.0007
4.098
40 cm
20
22.2
44.17
38.60
2.90E-07
30 cm
20
30.7
124.15
51.55
1.40E-08
20 cm
20
29.55
146.9
81.41
5.49E-11
4.098
Yes
10 cm
20
26.1
170.4
100.18
3.32E-12
4.098
Yes
2 cm
20
26.15
206.85
207.86
Yes
ip t
Significant Variance
cr
Yes Yes
4.098
Yes
an
us
4.098
4.098
F
P-Value
F crit
Significant Variance
87.1
38.95
2.65E-07
4.098
Yes
99.25
24.68
1.46E-05
4.098
Yes
5.44E-17
M
Table 4 - Visible photography ANOVA results for pooled (across burns) mean values of blood and soot pixel brightness.
Mean Blood Value
50 cm
20
48.65
40 cm
20
59.05
Mean Soot Value
d
n
Ac ce p
te
Sample Height
30 cm
20
65.15
121.5
37.58
3.77E-07
4.098
Yes
20 cm
20
66.4
139.4
61.13
1.97E-09
4.098
Yes
10 cm
20
63.3
168.45
126.31
1.26E-13
4.098
Yes
2 cm
20
55.25
172.85
168.83
1.47E-15
4.098
Yes
Table 5 - Infrared photography ANOVA results for pooled (across burns) mean values of blood and soot pixel brightness.
29
Page 30 of 31
Highlights
• •
Ac ce p
te
d
M
an
us
cr
•
Offenders of violent crime may attempt to conceal physical evidence by setting fire to the crime scene. Bloodstain evidence situated within the hot layers of a fire can be concealed by heavy depositions of dense soot. Longer wavelengths reduce the relative optical density of overlaying soot allowing for greater penetration into the material. Reflected infrared photography may visualise latent bloodstains cover by a heavy deposit of soot.
ip t
•
Page 31 of 31
S0379-0738(18)30660-1 https://doi.org/doi:10.1016/j.forsciint.2019.109874 FSI 109874
To appear in:
FSI
Received date: Revised date: Accepted date:
28 August 2018 29 June 2019 23 July 2019
Please cite this article as: B. Bastide, G. Porter, A. Renshaw, Detection of Latent Bloodstains at Fire Scenes Using Reflected Infrared Photography, Forensic Science International (2019), https://doi.org/10.1016/j.forsciint.2019.109874 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.
Detection of Latent Bloodstains at Fire Scenes Using Reflected Infrared Photography
Western Sydney University, School of Science & Health, Richmond, Australia.
us
a
cr
ip t
Belinda Bastidea*, Glenn Porterb, Adrian Renshawc
University of New England, School of Humanities, Arts & Social Sciences, Armidale, Australia.
c
Western Sydney University, Science of Science & Health, Richmond, Australia.
Ac ce pt e
d
M
an
b
Declaration of Interest: The are no interests to declare for this research.
Original Research Article 1
Page 1 of 31
Abstract Bloodstain evidence is an element of crime scene investigation often found at
ip t
scenes involving violence. Setting fire to the scene is a method sometimes used by offenders of crime in an attempt to conceal evidence. Fire often produces thick soot as a
cr
by-product of the combustion and has the potential to cover bloodstain patterns
us
rendering them latent. There is limited published material offering a method of detecting bloodstains hidden beneath dense soot deposits caused by fire. This project employed a
an
modified digital single-lens reflex (SLR) camera to investigate the application of reflected infrared photography to detect latent bloodstain evidence beneath varying deposited
M
overlaying soot densities. The potential of this technique was examined by photographing blood samples beneath soot from a scaled fire simulation. A qualitative
te
d
evaluation was completed by comparing images taken of a series of samples using both reflected infrared and standard visible light photography and corroborated with
Ac ce p
quantitative image analysis to support the findings. Results indicate that infrared photography can reveal latent bloodstains beneath a dense layer of soot in excess of ρ2.3 (550 nm) density with substantial clarity. The success of this technique is dependent on specific optical and specimen parameters. These parameters include i) the reflective properties of the background surface, ii) the spectral absorption properties of blood and iii) the ability of infrared wavelengths to transmit through the soot layer. Reflected infrared photography may provide crime scene examiners with a specialised field recording method that is easily executed and non-destructive to assist in visualising and locating latent bloodstain patterns beneath dense layers of soot. 1
Page 2 of 31
Keywords; infrared photography, bloodstains, latent evidence, forensic photography,
Ac ce p
te
d
M
an
us
cr
ip t
crime scene, blood evidence.
2
Page 3 of 31
Introduction Blood and bloodstain evidence recovered from scenes of crime involving violence is regularly used in forensic investigation for crime scene reconstruction and forensic
ip t
identification purposes [1-5]. A range of forensic science methods are employed to detect the presence of blood at sites involving violent crime. However, when bloodstain evidence is
cr
concealed due to the event of fire and covered by a deposit of acrid soot, detection using
us
more traditional trace blood detection techniques may not be effective [4].
Forensic aware offenders may employ methods of evidence concealment designed
an
to leave no trace evidence of the crime for forensic investigators to discover. Arson, or deliberate setting fire to property, can be a modus operandi to conceal evidence resulting
M
from violent crimes including homicide and has the potential to destroy physical and biological trace evidence. However, construction fires are often extinguished by fire
te
d
brigades before the structure and any trace evidence is entirely destroyed. Arson investigation is a multidisciplinary field requiring knowledge from a diverse
Ac ce p
range of different studies, including; i) chemistry (properties of accelerants and combustible material), ii) fire science (dynamics of fire), iii) crime scene investigation and iv) firefighting practices. This nexus of considerably different knowledge sources creates a high degree of complexity for forensic investigation that have involved fire. Smoke and sludge from a fire can cover and preserve bloodstain evidence on walls,
ceilings and other surfaces by establishing a protective layer [5]. There are a variety of optical and chemical techniques currently used within regular crime scene investigation and forensic photography to enhance bloodstain evidence at more conventional crime scenes [6] that usually do not involve fire. Chemical blood reagent techniques including; luminol, 3
Page 4 of 31
rhodamine 6G, amido black, leuco crystal violet, Hungarian red and various optical enhancement techniques are also employed [7]. However, excessive soot overlaying bloodstains may cause significant interference when chemical reagents are used in the
are often ineffective due to the carbon layer over the bloodstains.
ip t
recovery of this form of evidence [9] and non-destructive optical enhancement methods [6]
cr
Prior to the previous decade, any physical and biological remnants covered in soot
us
from fire were regarded as ineffectual [9-10]. It was often believed by crime scene examiners that the recovery of such evidence is not possible or too difficult. Consequently,
an
this situation has caused investigators to overlook the recovery of bloodstain evidence remaining in fire scenes [11]. This thinking was based on a suggestion that evidence exposed
M
to fire conditions would degrade and not permit more general investigation techniques to be effective or would not have any meaningful outcomes [9,11]. More recent studies
te
d
however, suggest various forensic techniques may be employed to visualise, enhance and collect fingermarks, DNA and blood evidence from fire affected sites including when the fire
Ac ce p
involved flashover [11-12]. The successful recovery of bloodstain evidence from fire scenes may lead investigators to better understand the events that occurred prior to the fire through bloodstain pattern analysis (BPA) and offer further identification evidence using DNA techniques from recovered trace blood. Photography is often considered as a primary source of evidence preservation for
several forms of physical evidence due to its non-destructive nature [6,13-14]. Infrared radiation consists of longer wavelengths [15] and has the ability to penetrate through some visually opaque medium resulting in latent material beneath becoming visible [6,16]. Infrared radiation also has the potential to discern between two different objects that 4
Page 5 of 31
appear optically similar under visible light due to the difference in spectral response from the materials irradiated by infrared radiation [16-17]. This metameric effect has significant potential value for forensic investigation at fire scenes.
ip t
When material is photographed using infrared radiation there are three possible recording outcomes; i) the material may absorb infrared radiation and become darker, ii)
cr
the material may reflect infrared radiation and become lighter or, iii) the material may
us
transmit infrared radiation and become transparent [18]. Infrared photography requires infrared sensitive film or digital cameras suitable (or modified) for recording infrared and a
an
light source containing infrared radiation [17-20].
When bloodstains are deposited onto various surfaces at crime scenes, they
M
immediately begin to oxidise or dry on surfaces. When blood dries it undergoes a change of colour from a vibrate red to a dark brown hue due to the chemical change as the oxidisation
te
d
process occurs. Cadd et al [21] explains the oxidisation of the haemoglobin changes composition from haemoglobin (Hb) to oxyhaemoglobin (HbO2) which then changes to
Ac ce p
metaemoglobin (metHb) and finally fully denatured into hemichrome (HC) [21]. Horecker [22] cites early experimentation work on the absorption of infrared radiation by haemoglobin and its derivatives by Merkelbach [23] as early as 1935. It was found that oxyhaemoglobin (HbO2) and methaemoglobin (metHb) can absorb a broad range of wavelengths, while carboxyhaemoglobin is unable to absorb infrared radiation [22]. Red blood cells are tightly packed with haemoglobin, consisting of approximately
one million haemoglobin molecules in one erythrocyte [24]. MetHb formation after heat exposure indicates the decomposition of blood components throughout an ageing process [7]. The exposure of blood to fire conditions can speed up this decomposition and ageing 5
Page 6 of 31
process, causing HbO2 material to be converted to metHb at an increased rate. This rate of decomposition is dependent on the exposure temperatures within the fire. Kuenstner & Norris [25] discovered the peak absorption for metHb and HbO2 occurs at approximately
ip t
630 nm and 930 nm respectively, with metHb absorbing approximately 10-times more infrared radiation than HbO2 [25]. The production of metHb from heat exposure and its
cr
greater infrared absorption properties may have a positive effect on the ability to record
us
latent bloodstains beneath soot when using reflected infrared photography.
Modelling the conditions of construction fires can provide an insight into the types of
an
evidence recovery techniques used by investigators within a forensic science context. The objective of conducting a scaled model experiment is not to replicate a full fire scenario,
M
rather it is to create easily controlled experimental fire conditions to determine the recoverability of bloodstain evidence using reflected infrared photography. The scaled
te
d
experiment will allow control parameters to be actively regulated including i) exposure temperature, ii) fuel load, iii) ventilation properties, iv) bloodstain location and v)
Ac ce p
determination of soot deposit density. The use of a quasi-steady fire model will profile accurate scaling of the burning rate and room environment produced by fire [26-27]. This model will provide replication of gas and wall temperature, gas species concentration and radiation flux at homologous locations when correct scaling laws are applied [27-28].
Materials & Methods
The experimental design consisted of several stages including; i) the fire modelling, ii) photographic recording and imaging, iii) measuring soot density at various hot layers and at differing wavelengths, iv) the examination of the images and v) analysis of results. 6
Page 7 of 31
Fire Modelling The scaled enclosure model in this project uses the theoretical design of a quasi-
ip t
steady model. This scaling scheme is used to replicate similar geometric and free burn behaviour from one scale to another [27]. The use of a quasi-steady model allows burning
cr
rate and gas temperature rise in realistic enclosure fires to be within those observed in free
us
burn cribs.
The enclosure model was designed from a full-scale bedroom with a single doorway
an
opening. The model was scaled to a 1:5 ratio of a full-scale building enclosure using plywood foundations. The lining of the enclosure was designed using plasterboard wall and ceiling
M
and metal tile flooring. The dimensions for both full scale and scaled enclosure buildings are shown in Table 1. The internal dimensions were measured whilst including the plasterboard
the scaled model.
te
d
wall linings, therefore small adjustments were made to allocate for plasterboard linings in
Ac ce p
This experiment uses a wooden crib fuel source and follows the burning regime using stick thickness. The stick dimensions used in the crib fuel are shown below in Table 2. The wood crib was constructed out of small clear pine wooden dowel, 2.5 cm in diameter and 400 kg/m3 density. Figure 1 shows the scaled modelled fire in action. This experiment used 200µL defibrinated horse blood sourced from a commercial
biological supplier, and was pipetted into specific locations of the gyprock on the rear wall of the scaled model. To place the bloodstains in the same location on the gyprock for each sample, a Perspex stencil was used as described by Farrar et al [20] (Figure 2). The locations of the 30 bloodstains were determined by drilling holes in the Perspex in a grid formation, 7
Page 8 of 31
parallel to the thermocouples places along the height of the enclosure. Each 200µL of blood was deposited onto each modelled fire wall using the grid for consistency of placement and location recovery. Figure 5 denotes the positioning of each bloodstain within different heat
cr
safe ignition of the fire and easy extinguishing if required.
an
us
ip t
layers. The scaled enclosed model was placed in an outdoor area on a flat hard surface for
te
Fuel Crib Ignition
d
M
walls>
To create replicable burning behaviour, the ignition of the fuel crib occurred in the
Ac ce p
same manner in each repetition. This allowed the same amount of heat energy to be released from the fuel. Ignition of the fuel crib was conducted using a kerosene-soaked rag. One piece of 200 mm x 200 mm cotton rag was soaked in 50 ml of kerosene and placed in the bottom of the fuel crib. The rag was then ignited using a hand-held fire starter gun. Once the fuel-soaked rag was ignited it then provided an initial heat flux to cause the wood dowel crib to ignite and become the dominate source of fuel source and heat energy.
8
Page 9 of 31
Fuel Crib Extinction A suffocation technique was used to extinguish the enclosure fire in a safe manner. As the fire was ventilation controlled, a sudden lack of oxygen caused fire to smother and
ip t
extinguish. This was achieved by closing the doorway opening and starving the fire of oxygen. Once the doorway was closed for a significant amount of time, (≈5 minutes), the
cr
door was reopened and a garden hose was used to spray water onto the crib. This small
us
amount of water caused any smouldering wood to become saturated. The hose was set on a low-pressure mist with a diameter small enough to cover the crib and not cause any
an
splashing onto the soot deposited walls within the enclosure.
M
Determining Soot Density
Transmission measurements recorded within this project examined the ability of
te
d
radiation to pass through a medium. Transmittance is the fraction of incident light at a specified wavelength that passes through a sample [29]. Percent transmission is expressed
Ac ce p
in equation form:
=
100
Equation 1: Percentage transmission.
The speed at which radiation wavelengths can pass through a medium depends upon the optical density of the material. The optical density of a material relates to the sluggish tendency of atoms to maintain the absorbed energy of an electromagnetic wave in the form 9
Page 10 of 31
of vibrating electrons before reemitting it as electromagnetic disturbance [29]. The more optically dense a material, the slower a wavelength will pass through it. The optical density
ip t
of a material is expressed as:
cr
= log
us
Equation 2: Optical density.
an
The density of the overlaying soot within the enclosure was determined using an Ocean
M
Optics HR2000 spectrophotometer. Small pieces of rectangular glass (microscope slides) were attached to the walls surface at the same heights where the bloodstains were located,
d
at 2 cm, 10 cm, 20 cm, 30 cm, 40 cm and 50 cm. Transmission measurements from these
te
slides at each height location were collected 5 times each and the mean transmission value determined. Transmission of each height was calculated using Equation 1 [29], then these
Ac ce p
measurements were converted into optical density measurements using Equation 2 [29]. The spectral transmission of the soot indicated transmission of visible light through to infrared. The longer infrared wavelengths were able to penetrate through the soot aggregate more efficiently than other wavelengths in the visible spectrum (see Figure 3) when optical densities were measured. A lower optical density result in the infrared wavelengths (900 nm) and in effect allows greater transmission in each of the fire layers (heights in the modelled fire). While the physical deposit of soot increases as the samples are placed in the higher heat layers of the fire, the optical density of soot also varies with
10
Page 11 of 31
different wavelength. Figure 3 also demonstrates the relationship between the relative density of the overlaying soot and wavelength.
ip t
us
cr
the scale enclosure.>
Camera Modifications and Spectral Sensitivity
an
The project used a modified Canon digital single-lens reflex (SLR) camera that was converted to record exclusively within the infrared region (>820 nm). The camera was
M
modified by removing the infrared blocking filter over the sensor and replacing it with an infrared transmission filter (equivalent to a Wratten 87C filter).
This longpass filter
te
d
transmits longer wavelengths above 820 nm while rejecting shorter wavelengths less than 820 nm. Figure 4 provides the spectral transmission of the optical filter fitted over the
Ac ce p
camera sensor.
Standardising Exposure Values and Camera Settings
11
Page 12 of 31
To enable a meaningful comparison between the visible light and infrared images, the images must be captured using exposures of equivalent exposure density or pixel brightness values. The difficulty when comparing images taken with different radiation
ip t
sources and camera spectral sensitivity is determining equivalent camera exposures for each radiation source that result in the same image density range. In other words, taking
cr
two different camera exposures using different radiations sources, different camera spectral
us
sensitivity and differing optical filtration, to achieve exposures that can be considered as alike. A calibration factor between the differing overall sensitivity of each sensor was
an
determined.
The ‘equivalent’ camera exposures were tested using a Kodak Q13 ‘Grey Scale’
M
reference standard with each camera (visible light and infrared) and using the same tungsten lighting with even illumination. Tungsten lighting was selected because it has a
te
d
spectral distribution characteristic that includes both visible light and infrared radiation. Testing exposure values consisted of taking a series of photographs of the subject with an
Ac ce p
attached Kodak Q13 ‘Grey Scale’ reference standard with each camera using a range of aperture settings.
The eyedropper tool in Adobe PhotoshopTM was used to measure the pixel
brightness of the ‘M’ segment of the Kodak Q13 greyscale reference. The ‘M’ segment has an 18% reflectance with an optical density value of ρ0.7. The eyedropper tool was used to measure the numerical value of pixel brightness between 0 and 255 for 8-bit images. The exposure values for each photographic method were chosen by measuring the ‘M’ segment and choosing the settings that gave the closest value to 18% grey at a 128-brightness value [20]. Table 3 provides the results of the camera exposure calculations for each type of 12
Page 13 of 31
photography. All photography throughout the experimental work was taken at this novel standardised exposure method and the lighting remained consistent during the calibration
ip t
test.
cr
us
Visual Examination and Image Analysis
Visual examination of the test photographs was conducted on a calibrated computer
an
monitor. The visual examination purpose was to determine the soot density parameters where the latent bloodstains became visible. The visual examination was completed for all
M
photographs taken using both reflected infrared and standard visible light photography. Image analysis used a combination of Adobe PhotoshopTM and ImageJ TM software to
d
quantify the effectiveness of infrared photography as a technique to visualise latent
te
bloodstains beneath the soot layers. PhotoshopTM was used to prepare the photographs for
Ac ce p
analysis and ImageJ was used to generate pixel brightness values from each image. Data from the image analysis was compiled on all experimental and control photographs using both the standard and infrared cameras. Each photograph was examined in Photoshop and using black tape outlining the
sample plasterboard as a guide, each image was cropped and resized to 2500 x 2077 pixels. The images remained in an upright orientation as indicated by the small ‘North’ indicator placed on each sample. The image was then converted into an 8-bit image and all colour information was discarded to maintain pixel brightness consistency throughout the analysis. This was completed by selecting Image > Mode and by selecting both ‘Greyscale’ and ‘8 13
Page 14 of 31
Bits/channel’. The image was then saved by using the File > Save As function. All images were saved in JPEG format and were set to maximum quality. Each image was opened in ImageJ and the line tool was selected to draw a linear line
ip t
in a horizontal orientation across the pixels where the blood drops were located to ensure data from all pixels across the width of the image was generated. By selecting Analyse > Plot
cr
Profile, a plot profile line was graphed by the software and the pixel brightness values were
us
saved into a spreadsheet. Further readings were taken for each sample repeat by drawing vertical lines throughout the image where the blood droplets were located. This generated
an
11 individual pixel brightness readings where the bloodstains were located. The plot profile is generated from the pixel brightness values contained within the image. Each pixel is
M
assigned a grey value between 0 and 255 representing the grey scale of each individual
Experimental Controls
te
d
pixel.
Ac ce p
Control measurements were taken of the pixel brightness values of the background white paint and the bloodstains to compare the spectral difference between the two objects and determine if there was significant difference between the two. This was conducted using both visible and infrared images captured of each of the samples. The images were examined in Adobe PhotoshopTM using the eyedropper tool to determine the pixel brightness of both the background and the blood. These measurements were then used in the analysis including an ANOVA evaluation to statistically represent any variance between the visible and infrared images [8]. Despite using very specific burn modelling controls including replica fuel loads and ignition processes, fires may still produce variations in burn 14
Page 15 of 31
patterns and soot deposition. This project used four model fires with five samples in each burn layer resulting in 20 samples per burn. The statistical modelling used a mean value for
ip t
all model fire replications to counter burn pattern variability.
Results
cr
Experimental samples were photographed using reflected infrared and standard
us
visible light photography techniques using the calibrated exposure with consistent illumination and relative exposure values. Several elements were examined to determine
an
whether infrared was an effective forensic imaging method to visualise bloodstains beneath dense soot deposits resulting from a fire. The results were evaluated using the following
i)
M
information and data;
A qualitative assessment using a comparison of images viewed on a
ii)
te
d
calibrated monitor,
Consideration of the relative density of the soot on each heat layer
Ac ce p
(expressed as height in the scale-model burn), iii)
A quantitative assessment of the difference between the pixel brightness readings between the soot and blood+soot deposits at each heat layer or height (also considered as contrast between the two areas).
The four replicants produced consistent results for burns 1,2 and 4. However burn 3
resulted in significantly less soot deposit due to the unpredictable nature of fire. Nevertheless, the other three burns provided consistent results to draw conclusions while the third burn may be considered as an outliner. The only burn height that was of interest to
15
Page 16 of 31
examine whether infrared was more effective than standard visible light was the 50 cm scale-model height due to the greater deposit of soot in the hotter layer. The examination of the relative soot densities at different wavelengths indicated
ip t
there is a significant difference for the ability of differing wavelengths to penetrate through soot deposits. Figure 3 indicates several elements to support this outcome; i) soot density
cr
increases within the hotter (higher) fire layers and, ii) longer wavelengths lower the relative
us
density which results in greater potential penetration capacity. Figure 3 reveals a relative optical density at 50 cm height of ρ2.9 for 400 nm wavelength, ρ2.3 at 550 nm, ρ1.8 at 700
an
nm and ρ1.3 at 900 nm. This result suggests infrared radiation has an improved potential to image through soot deposit layers than visible light. All other fire layer heights have a similar
M
trend however these heights are less important to this project.
Since relative density is a logarithmic value the difference displayed in Figure 3 are
te
d
significant. At 50 cm height the difference between 400 and 900 nm wavelengths are more than 30-times greater, while between 550 nm approximately 8-times relative density
Ac ce p
difference. This is a significant result when considering the potential of infrared photography’s ability to penetrate the covering soot deposit to record underlying bloodstains.
The visual assessment of standard visual light photography indicated it was difficult
to see the bloodstains with soot deposits >ρ1.5 relative density (550 nm) while infrared photography results show a substantial improvement in the visualisation of bloodstains at increased relative density levels (>ρ2.3 at 550 nm which is approximately 8-times greater). Further quantitative assessment of the visible light and infrared images using pixel brightness values (0-255 range) measured directly from each image also indicated 16
Page 17 of 31
improvement in the visualisation of bloodstains at higher relative densities of overlaying soot. Figure 6 plots the pixel brightness values from each burn at differing heights (fire hot layers). These plots display the difference between the brightness value (using a calibrated
ip t
exposure as indicated in the methods section between each imaging system) of the soot only to the pixel value of the soot+blood. This difference is considered as contrast within the
us
range within silver halide or film photographic technology.
cr
photographic recording capacity and is analogous to Dmax-Dmin expression of contrast
Figure 6 plots, with the exclusion of Burn 3 considered as an outlier, demonstrates
an
no advantage for infrared imaging at the lower heights and density deposit sections. This is to be somewhat expected because infrared photography is generally a lower contrast
M
recording method than standard white light photography, for general usage. This is certainly the case when comparing the contrast difference between the visible light and infrared
te
d
photography at those lower density coverage heights. However, as seen at the 50 cm height (and to some degree 40 cm height too), there is a significant change in contrast towards an
Ac ce p
improvement of contrast for the infrared recording. Greater pixel difference or contrast produces greater visibility. The pixel brightness values indicate a significant improvement in visualisation of blood under soot density >ρ2.3 (550 nm).
17
Page 18 of 31
ip t
different hot layers. Note; burn 3 is an outlier.
Single-Factor Analysis of Variance
cr
The results of ANOVAs performed on the mean pixel brightness data to determine
us
whether there were any significant differences (p<0.05) between the blood and soot values for visible and infrared images are displayed in Tables 4 and 5. These results indicated that
an
significant variation between the blood and soot values occurred for all heights within the enclosure model. However, the calculated F value was much greater than the critical value
M
for the infrared images at 50 cm height. This indicates a greater magnitude of variance between the blood+soot and soot recorded brightness when captured using reflected
te
d
infrared radiation photography. The ANOVA results further supports the previous findings.
Ac ce p
Discussion
Qualitative and quantitative analysis indicated that when soot density increased, the
difference between standard visible light photography and reflected infrared photography was affected. At the greatest soot density in the upper hot layers (50 cm height) of the scaled-model fire, bloodstains photographed with standard visible light photography methods failed to record the bloodstains beneath the soot with clarity. However, bloodstains photographed using reflected infrared photography at soot density >p2.3 (550 18
Page 19 of 31
nm) significantly improved the visualisation of the bloodstains. Soot deposits ≥ρ1.5 were most effective at concealing bloodstains using standard visible light photography. These soot deposits were located in the hot layer of the scaled enclosure (40 - 50 cm height).
ip t
Soot deposits between ρ0 and ρ1.0 density were the least effective at concealing bloodstains using standard visible photography. These deposits were located in the lower
cr
region of the scaled enclosure (0 - 20 cm height). Experimental work involving measuring
us
the relative optical density of overlaying soot indicated longer wavelengths lowered the relative optical density of overlaying soot. This means that longer wavelengths like infrared
an
has greater potential to penetrate through the material and record substances like blood underneath the soot. In similar fashion to Farrar et al [20] description of how blood is
M
visualised underneath painted surfaces, the visualising of blood beneath soot requires the radiation source (infrared) to transmit through the overlaying soot, be absorbed by the
te
d
bloodstain rendering it darker [25] and then reflect off the wall background to produce contrast resulting in improved visualisation. With dense soot deposits >ρ2.5 wavelengths
Ac ce p
within the visual spectrum will have a reduced ability to penetrate through the overlaying soot resulting in ineffectual visualisation of bloodstains and possibly remaining latent. Infrared photography combines both transparent [18] properties of the soot and absorption of the blood [25] making it a more suitable option for dense overlaying soot. Statistical analysis using ANOVA compared mean pixel brightness values of
blood+soot and soot from standard visible light and reflected infrared images. Reflected infrared photography exhibited significant variance between the tonal values for blood and soot for the higher soot densities. These results provided further corroborate the findings
19
Page 20 of 31
that reflected infrared photography has the potential detect latent bloodstains beneath overlaying soot from arson scenes. The scaled enclosure model used in this project was designed to replicate fire
ip t
behaviour and conditions experienced in full-scale scenarios. The engineered design of the model ensured a broad soot density gradient was achieved such that bloodstains could be
cr
concealed. This included using a calculated, easily combustible fuel source and providing
us
adequate ventilation for control of combustion. Furthermore, modelling fire behaviour provides experimental replication and a model is easy to construct with the availability of
an
materials and inexpensive cost. Using a scaled enclosure built with strong foundations to withstand multiple fire simulations allowed the investigation to achieve results that can be
M
effectively applied to full scale fire scenarios.
The results from this work are relevant to the current forensic field as they provide
te
d
crime scene examiners with a method of detecting bloodstains situated beneath soot. Other techniques investigated for this type of evidence recovery include traditional techniques as
Ac ce p
indicated by Tontarski et al., [9]. The use of traditional techniques was successful in detecting bloodstains beneath soot, however interactions between the chemicals and soot caused difficulties. Results obtained within this work indicate that it is also possible to photograph bloodstains beneath dense overlaying soot using non-destructive reflected infrared photography method.
Conclusion Reflected infrared photography was successful in detecting latent bloodstains beneath dense overlaying soot. This success was achieved by combining the spectral 20
Page 21 of 31
responses of four parameters necessary for this technique including; i) the increased infrared absorption properties of blood after exposure to heat, ii) the infrared transmission and penetration of soot layer, iii) the infrared reflection of the painted background and iv)
ip t
the sensitivity of the digital camera and spectral output of the light source. The combination of these individual infrared responses of blood, soot and background, causes a difference in
cr
tonal values or contrast recorded between bloodstains and overlaying soot, hence latent
us
bloodstains may become visible.
The qualitative comparison between images using standard visible light photography
an
and reflected infrared photography revealed the infrared techniques were able to detect bloodstains beneath dense overlaying soot (and penetrate the soot layer). This qualitative
M
result was confirmed by further quantitative and statistical analysis. Image analysis was used to generate pixel brightness profiles of the images and graphically demonstrate the
te
d
ability to visualise the clarity of bloodstains once increasing soot densities were applied. The difference in mean pixel brightness between blood+soot and soot deposits were verified
Ac ce p
using ANOVA statistical analysis that indicated a greater variance and visualisation in images using reflected infrared photography at more dense soot coverage. The overall results indicated standard photography techniques could no longer
visualise bloodstains after soot densities of ρ1.5 (550 nm) were achieved. Reflected infrared photography was able to produce effective clarity and visualisation of bloodstains beneath soot density in excess of ρ2.3 (550 nm). Since density is a logarithmic scale, these differences are significant.
References 21
Page 22 of 31
[1] Cadd S., Li B., Beveridge P., O’Hare W.T., Campbell A., Islam M., (2016) The non-contact detection and identification of blood stained fingerprints using visible wavelength reflectance hyperspectral imaging: Part 1. Sci Justice. 56 181-190.
ip t
http://dx.doi.org/10.1016/j.scijus.2016.01.004 [2] Cadd S., Li B., Beveridge P., O’Hare W.T., Campbell A., Islam M., (2016) The non-contact
us
reflectance hyperspectral imaging: Part 2. Sci Justice. 56 191-200.
cr
detection and identification of blood stained fingerprints using visible wavelength
an
http://dx.doi.org/10.1016/j.scijus.2016.01.005
[3] Barni, F., Lewis, S., Berti, A., Miskelly, G., Lago, G., (2007) Forensic application of the
M
luminol reaction as a presumptive test for latent blood detection. Talanta. 72(3) 896-913. https://doi.org/10.1016/j.talanta.2006.12.045
d
[4] Vandenberg, N. Oorschot, R.A., (2006) The use of Polilight® in the detection of seminal
te
fluid, saliva and bloodstains and comparison with conventional chemical-based screening
Ac ce p
tests. J Forensic Sci. 51(2) 361-70. https://doi.org/10.1111/j.1556-4029.2006.00065.x [5] Redsicker D., O’Connor J., Practical fire and arson investigation. CRC Press, Boca Raton, 1997.
[6] Porter G., Forensic Photography in Rennie J., et al (Eds) McGraw-Hill Yearbook of Science & Technology. McGraw-Hill, New York 2014 pp124-129. [7] James S.H., Kish P.E., Sutton P.T., Principles of bloodstain pattern analysis: theory and practice. CRC Press, Boca Raton, 2005. [8] Sokal R.R., Rohlf F.J., Biometry: The principles and practices of statistics in biological research, 3rd Edition. W. H. Freeman, 1994.
22
Page 23 of 31
[9] Tontarski K.L., Hoskins K.A., Watkins T.G., Brun-Conti L., Michaud A.L., (2009) Chemical enhancement techniques of bloodstain patterns and DNA recovery after fire exposure. J Forensic Sci. 54(1) 37-48. https://doi.org/10.1111/j.1556-4029.2008.00904.x
ip t
[10] Bradshaw G., Bleay S., Deans J., NicDead N., (2008) Recovery of fingerprints from arson scenes: Part 1 - latent fingerprints, J Forensic Ident. 58(1) 54-82. ISSN: 0895173X.
cr
[11] Deans J., (2006) Recovery of fingerprints from fire scenes and associated evidence. Sci
us
Justice. 46(3) 153-168. https://doi.org/10.1016/S1355-0306(06)71589-1.
[12] Moore J., Bleay S., Deans J., NicDead N., (2008) Recovery of fingerprints from arson
an
scenes: Part 2 - fingerprints in blood. J Forensic Ident. 58(1) 83-108. ISSN 0895173X. [13] Porter G., (2011) A new theoretical framework regarding the application and reliability
M
of photographic evidence. IJE&P, 15(1) 26-61. https://doi.org/10.1350/ijep.2011.15.1.367. [14] Porter G., (2007) Visual culture in forensic science. Aust J For Sci. 39(2) 81-91.
te
d
https://doi.org/10.1080/00450610701650054 [15] Robinson, E., Crime scene photography. Academic Press, Burlington 2007.
Ac ce p
[16] McKechnie M.L., Porter G., Langlois N., (2008) The detection of latent residue tattoo ink pigments in skin using invisible radiation photography. Aust J For Sci. 40(1) 65-72. https://doi.org/10.1080/00450610802047580 [17] Denny K.A., (2015) Recognising changes in visual representation of clothing in CCTV imaging. J Criminological Research, Policy & Practice. 1(4) 233-238. https://doi.org/10.1108/JCRPP-08-2015-0033 [18] Porter, G., Photography & optical enhancement of physical evidence. in Jamieson A. & Moenssens A. (eds.), Wiley encyclopedia of forensic science. 2nd edition. Wiley & Sons. New York. 2013. 23
Page 24 of 31
[19] Verhoeven G., (2008) Imaging the invisible using modified digital still cameras for straightforward and low-cost archaeological near-infrared photography. J Archaeol Sci. 35(12) 3087-3100. https://doi.org/10.1016/j.jas.2008.06.012.
ip t
[20] Farrar A., Porter G., Renshaw R., (2012) Detection of Latent Bloodstain Patterns Beneath Paint Using Reflected Infrared Photography, J Forensic Sci. 57(5) 1190-1198.
cr
https://doi.org/10.1111/j.1556-4029.2012.02231.x
us
[21] Cadd S., Li B., Beveridge P., O’Hare W.T., Campbell A., Islam M., (2018) Age
determination of blood stained fingerprints using visible wavelength reflectance
an
hyperspectral imaging. J.Imaging. 4 (12) 141 1-11. doi:10.3390/jimaging4120141 [22] Horecker B. L., (1943) The absorption spectra of haemoglobin and its derivatives in the
M
visible and near infrared regions. J Biol Chem. 148 173-183.
[23] Merkelbach O., (1935) Infrared absorption and infrared photography of normal and
te
d
carbon monoxide (luminous gas) poisoned blood. Schweiz Med Wschr, 65 1142. [24] Botonjic-Sehic E., Brown C.W., Lamontange M., Tsaparikos M., (2009) Forensic
Ac ce p
application of near-infrared spectroscopy: Aging in bloodstains. Spectroscopy. 24(2) 42,4448. ISSN: 08876703
[25] Kuenstner J.T., Norris, K. H. (1994) Spectrophotometry of human hemoglobin in the near infrared region from 1000 to 2500 nm. J Near Infrared Spec. 2(2) 59-65. https://doi.org/10.1255/jnirs.32
[26] Almirall J., Furton K.G., Analysis and interpretation of fire scene evidence. CRC Press, Boco Raton. 2004. [27] Croce P.A., Xin Y. (2005) Scale modeling of quasi-steady wood crib fire in enclosures. Fire Safety J. 40(3) 245-266. https://doi.org/10.1016/j.firesaf.2004.12.002 24
Page 25 of 31
[28] Jolly S., Saito K. (1992) Scale modelling of fires with emphasis on room flashover phenomenon. Fire Safety J. 18(2) 139-182. https://doi.org/10.1016/0379-7112(92)90036-C [29] Jacobson R.E., Ray S.F., Attridge G.G., Axford N.R., The manual of photography. 9th
Ac ce p
te
d
M
an
us
cr
ip t
Edition. Focal Press, Oxford. 2000.
25
Page 26 of 31
1:5
1:5 + Plasterboard*
Length of Room (mm)
3130
626
646
Width of Room (mm)
3230
646
2560
512
830
166
2030
406
666
cr
532
us
166 406
an
Height of Room (mm) Width of Doorway(mm) Height of Doorway(mm)
ip t
Full Scale
Ac ce p
te
d
M
Table 1 - Dimensions of full scale and scaled enclosure model (*measurements for outside perimeter + 20 mm width of plasterboard lining of the model).
26
Page 27 of 31
Dimensions 150 mm
Width þ
25 mm
Layers N
10
Sticks per layer n
3
cr
ip t
Length l
Ac ce p
te
d
M
an
us
Table 2 - Stick dimensions used in crib fuel source.
27
Page 28 of 31
Lens
50 mm Macro
50 mm Macro
ƒ/stop
8.0
8.0
Shutter speed
1/20th sec
1/125th sec
ISO
100
100
White Balance
Custom
Custom
U distance
2.0 m
2.0 m
EV (Exposure Value)
10.3
13
Light source
Tungsten
ip t
Visible Camera
us
cr
IR Camera
an
Tungsten
Ac ce p
te
d
M
Table 3 – Calibrated camera settings used throughout the experiment
28
Page 29 of 31
Sample Height
n
Mean Blood Value
Mean Soot Value
F
P-Value
F crit
50 cm
20
20.3
51.85
13.41
0.0007
4.098
40 cm
20
22.2
44.17
38.60
2.90E-07
30 cm
20
30.7
124.15
51.55
1.40E-08
20 cm
20
29.55
146.9
81.41
5.49E-11
4.098
Yes
10 cm
20
26.1
170.4
100.18
3.32E-12
4.098
Yes
2 cm
20
26.15
206.85
207.86
Yes
ip t
Significant Variance
cr
Yes Yes
4.098
Yes
an
us
4.098
4.098
F
P-Value
F crit
Significant Variance
87.1
38.95
2.65E-07
4.098
Yes
99.25
24.68
1.46E-05
4.098
Yes
5.44E-17
M
Table 4 - Visible photography ANOVA results for pooled (across burns) mean values of blood and soot pixel brightness.
Mean Blood Value
50 cm
20
48.65
40 cm
20
59.05
Mean Soot Value
d
n
Ac ce p
te
Sample Height
30 cm
20
65.15
121.5
37.58
3.77E-07
4.098
Yes
20 cm
20
66.4
139.4
61.13
1.97E-09
4.098
Yes
10 cm
20
63.3
168.45
126.31
1.26E-13
4.098
Yes
2 cm
20
55.25
172.85
168.83
1.47E-15
4.098
Yes
Table 5 - Infrared photography ANOVA results for pooled (across burns) mean values of blood and soot pixel brightness.
29
Page 30 of 31
Highlights
• •
Ac ce p
te
d
M
an
us
cr
•
Offenders of violent crime may attempt to conceal physical evidence by setting fire to the crime scene. Bloodstain evidence situated within the hot layers of a fire can be concealed by heavy depositions of dense soot. Longer wavelengths reduce the relative optical density of overlaying soot allowing for greater penetration into the material. Reflected infrared photography may visualise latent bloodstains cover by a heavy deposit of soot.
ip t
•
Page 31 of 31