Assessing iPhone LiDAR & Recon‐3D for determining area of origin in bloodstain pattern analysis

Bloodstain pattern analysis (BPA) has proven to be a useful tool in forensic and criminal investigations for quite some time. Traditionally, documenting a crime scene for a bloodletting event was completed using manual techniques, physical strings, and a tape measure. In more recent years, laser scanners and 3D software programs have become a preferred method to capture accurate data that improves the validity and reliability of BPA. The initial cost of laser scanning equipment is relatively high, rendering these systems inaccessible to some police and smaller agencies. Recon‐3D is a newly developed iPhone application that utilizes the iPhone LiDAR sensor in combination with video data to create 3D point clouds of crime scenes. To assess the viability of Recon‐3D for area of origin analysis, two tests were performed. One was a series of bloodstain impacts which were analyzed in FARO Zone 3D software, while the second was a series of 6 repeated Recon‐3D scans of two 90‐degree walls which was then compared to the FARO Focus S350 scanner using CloudCompare software. A total of eight impact patterns were made at three different distances from a wall. The area of origin was measured and compared to the known location of the blood source. The average total 3D error for the area of origin set at 25, 50, and 100 cm from two perpendicular walls was found to be 6.04, 15.16, and 36.59 cm, respectively. These results are similar to past studies where programs such as HemoSpat have been used. The results of the point cloud comparison show that on average, 95% of the points from Recon‐3D fall below a threshold of 3.6 mm when compared to a FARO Focus S350 laser scanner. Thus, the results of this test suggest that Recon‐3D is an accurate and affordable scanning application for bloodstain patterns at crime scenes and the data provide acceptable results for area of origin analysis in BPA programs which accept laser scanner data.


| INTRODUC TI ON
The field of Bloodstain Pattern Analysis (BPA) provides investigators with essential insight into bloodletting crime scenes.Specialists in this field examine blood present at a crime scene to provide understanding as to the sequence of actions, location of people, and the mechanisms of a crime [1,2].An expert in this field gathers information from the crime scene and reconstructs the actions of bloodletting events with validated methodologies.Expert opinions are rendered by specialists who apply validated documentation techniques and measurements while considering the physical behavior of blood [1,2].The biological and physical makeup of blood is one of the many factors affecting how bloodstain patterns are created in different environments.Blood is a complex fluid that is affected by the surrounding physical and atmospheric conditions.The unique composition of blood contributes to its physical properties including density, viscosity, and surface tension [3].
BPA experts characterize different types of bloodstains by the mechanism in which the stains were created although there is a trend to move away from a mechanistic approach to naming bloodstain patterns.Impact patterns are of particular interest to forensic investigators as they provide valuable insight necessary for scene reconstruction [4].Impact patterns are created when an object strikes a source of liquid blood, most often resulting in a V-shaped pattern.
Analysts study the characteristics of impact patterns to calculate the Area of Origin (AO) and area of convergence to determine the location of a blood source at the time the impact occurred [1].The traditional accepted practices and standardized workflow have been extensively documented by the Scientific Working Group for Blood Stain Pattern Analysis (SWGSTAIN) and are outlined below [5].
The initial step in the investigative BPA workflow is a visual examination of the scene and choosing bloodstains suitable for analysis.The analyst will then assign numbers, and place scales or markers next to clusters or areas of interest [6].A variety of individual bloodstains are selected from either side of the blood pattern, and stain selection is guided by validated protocols to improve the accuracy and height estimation of blood sources [1,6].The selected stains are individually photographed and sized based on the ellipse fitting methods to determine the alpha angle (impact angle) and gamma angle (directional trajectory) [6,7].The angle of impact is calculated using a mathematical relationship represented by the following [6][7][8]: where w and l represent the width and length of the bloodstain ellipse.
Although BPA has been utilized in police work as early as the 1920s, the physical stringing method was not developed until the 1970s [9].Bloodstains were selected by investigators, documented, and affixed with strings to determine the area of convergence and the area of origin of the blood source (Figure 1).Typically, these practices were conducted by means of physical strings and measuring tape, leading to complications at the scene such as contamination, tediousness, and subjectivity [8].Software programs such as BackTrack and HemoSpat were created to virtually recreate the stringing method with user-friendly software applications [8,10].In turn, BPA analysis continued to increase in speed, quality, and reliability in line with the 2009 National Research Council recommendations [11].The emergence of these newly validated computerized programs reduced the time of processing, subjectivity, and human error, and decreased the number of investigators required to be present at the scene [12].

| 3D technology: Photogrammetry, laser scanning, and LiDAR
In BPA, Three-Dimensional (3D) techniques may be applied to assist • Comparison of Recon-3D to industry standard laser scanning systems.
• Inexpensive and accessible software to document blood spatter with spatial information in a timely manner.

F I G U R E 1
Example of physical stringing method.The square box highlights the area from where the impact originated.
Photogrammetry is a technology that allows the user to draw reliable, accurate information about a scene through a combination of measuring, recording, and analyzing photographic data.Images are captured of an object by taking photographs from multiple perspectives with overlapping portions of the photographs.The object can then be reconstructed in 3D, and measurements can be made.[17].In 2019, a focus group was conducted consisting of forensic investigators to gain insight into the cost-benefit analysis of using 3D technology and laser scanning for crime scene investigations.Although most participants unanimously recognized the benefit for law enforcement agencies, the initial investment of this technology makes it inaccessible for most organizations [17].

| iPhone LiDAR sensor
A release by Apple Computers in the fall of 2020 introduced the world to the first smartphone with a built-in LiDAR system [20].The laser emits a grid of 12 × 12 infrared dots which then shift position to create an overall larger pattern of dots (Figure 2).These individual laser dots bounce off surrounding objects and surfaces, ultimately returning to the sensor [20].Developers may access these data to create a 3D model based on the depth map and photographic information [21].Although only a few years old, several studies have examined the application of the iPhone LiDAR scanning technology to forensic investigations.Human body measurements including autopsy documentation by forensic pathologists tested the iPhone LiDAR scanner with success [14,20,21].Maiese et al. [21] concluded several notable advantages including ease of operation, no specialized skills necessary, affordability, and improved jury comprehension during trials [22].

| Recon-3D
Recon-3D (R3D) was released in May 2022 and is a newly developed R3D allows the users to accurately scale their scan data with target detection from AprilTags that are placed and measured within a scene.AprilTags act as reference targets within the scanning space to enable the user to accurately scale their point cloud data within a 3D environment [23].The distance between the AprilTags can be measured manually or with a laser distometer and then input into the Target Distance section of the scan settings.Although this is an optional step, it provides the user with a necessary reference measurement to accurately scale their data and is recommended for forensic uses (Figure 3).
R3D also allows the user to choose the resolution of the scan, with settings in the range of 1 to 30 mm, with 1 mm being the finest setting.The user may adjust the resolution setting based on the amount of detail that is required for the output scan (Figure 4). in the e57 file format.The data can then be used in 3D software such as CloudCompare, Map360, or FARO Zone 3D.

| PURP OS E
In this study, the quality and accuracy of the R3D application and iPhone LiDAR data for impact patterns were evaluated based on a comparison between ground truth values and calculated AO values.
The quality of R3D scans was assessed on based reproducing similar results, and accuracy was assessed and compared to known results, such as FARO scanning data.The analysis was conducted with R3D scan data that were analyzed in FARO Zone 3D with the Blood Spatter power tool plugin.

| Impact testing
A comparative research model similar to previous BPA studies was used as a framework for the experimental setup [6,8,16].A custombuilt impact rig was set up at three distances from two perpendicular walls, 25, 50, and 100 cm in the corner of the testing room (Figure 5).
The placement of the rig in the corner of the room was chosen to make the analysis include two walls instead of only one for each impact [24].
3D-printed square scales with known inner dimensions (10 × 10 cm, 20 × 20 cm) were created with built-in bubble levels using a Creality 3D printer (Figure 6).This was done to improve any lens distortion and perspective issues with the photographs, which could be fixed with Adobe Photoshop in the post-processing workflow.
The 3D-printed square scales were painted black to create contrast from the white test wall.For future research, it is recommended that the scales should be initially printed with a dark-colored filament.This will minimize the need to paint them after printing and decrease the chance of warping or deformation.
The initial experimental setup, including the placement and measurement of two AprilTags, was conducted to ensure accurate recording of initial experimental conditions and ground truth values (Figure 7).AprilTags were placed within a scanning scene to accurately scale the point cloud in 3D space.The distance between the two targets was recorded for input into the scaling options offered within the Recon-3D application.For each experimental distance, the impact rig was placed at an equal distance between both perpendicular walls, and measurements were taken from the center of the impact rig target location from both the left and right walls with a steel tape measure.
Once the rig was placed accurately, the height of the base was measured to record the ground truth value for the Z coordinate, where the striking surface was located on the base of the rig.A level was applied on all dimensions of the drop tube enclosing the striking piston before each impact to ensure it was released from the tube evenly from all directions.To determine the appropriate amount of Once all the initial conditions and measurements of the ground truth were recorded, approximately 2 mL of blood was prepared in a clean syringe and loaded onto the striking surface of the impact rig (Figure 8).The impact rig operates by pulling a pin attached to a long string, thereby releasing the piston to drop freely onto the blood on the striking surface of the rig.After the impact was created, the blood was allowed to dry for approximately 10 min with the aid of an oscillating fan.The impact was then inspected for usable clusters in at least three locations of the overall pattern.
3D-printed square scales (10 × 10 cm and 20 × 20 cm) equipped with bubble levels were placed around clusters of useful stains and were chosen based on previous research suggestions and recommendations (Figure 9).For reliable results, it was important that well-formed symmetrical ellipses indicating directionality were selected from various locations in the scene, aiming to capture a wide spread of stains within each selected cluster [5,25].After all the square scales were placed and secured with tape, the bloodstains were photographed using a Nikon D7100 camera with a 1:1 60 mm macro lens (Figure 10).The selection of the lens was important to limit perspective and distortion issues in post-processing.A high-resolution, JPEG photograph of each cluster was taken at a 90-degree angle to the surface of the stains and subsequently uploaded to a computer to assess for lighting and focus before moving on to the final step of the scan data collection.
After the photographs were assessed, the whole scene was in each scene (Figure 13).
With the photographs accurately placed onto the point cloud, the next action taken was to complete ellipse drawing.Resources were consulted to accurately trace the ellipses for the best results.
It has been noted that tiny spatter droplets rendered inaccurate trajectories and were therefore eliminated from the analysis [4,6].The ellipse tracing process was repeated on each cluster (three to five clusters per impact, depending on the individual impact patterns), choosing only well-formed ellipses traveling in the same general direction (Figure 14).Once FARO Zone 3D had at least two photographs with ellipses drawn, it automatically calculated the estimated AO [19,25,26].
The ellipse tracing method was completed for suitable stains for each photo for all impacts.Following ellipse tracing, a single point was created with coordinates matching the known location coordinates of the blood source on the striking surface of the impact rig.

| Point cloud comparisons to FARO S350 laser scanner
In order to assess the contribution of error of Recon-3D in comparison to an established terrestrial laser scanner, a FARO Focus S350 scanner, which has been previously studied, was chosen as a reference.The same corner wall section used in the impact study was scanned by using the ¼ resolution and 4× quality setting on the laser scanner.This is a common setting at crime scenes and results in a point spacing of approximately 6 mm at 10 m.The data were processed in FARO Scene and then exported as an e57 file.
Similarly, Recon-3D was used to scan the same section of wall, six times in a similar manner as was done for the impact tests.A setting of 1 mm resolution and a scan time of approximately 60 s was maintained for each scan.The scans were conducted in a similar manner, using a similar scan path and distance from the wall.The Recon-3D data were then processed on the cloud, and the e57 output file was subsequently imported into CloudCompare.Each of the six scans was roughly aligned to the FARO S350 scan data by using the manual translation tools before the Fine Registration method could be used in CloudCompare.
The comparison to the FARO S350 scanner data using the Compute Cloud/Cloud Distance analysis tool in CloudCompare was computed for each respective Recon-3D scan, and the results were analyzed to determine how similar/different each scan was in relation to the FARO scanner data (Figure 15).

| RE SULTS
A comparative data analysis was chosen to determine if R3D has equivalent or similar capabilities to produce point cloud data similar to past studies where the FARO S350 laser scanner was utilized.To determine the relative total 3D distance taking into consideration the x, y, and z values at each impact, the following mathematical calculation for 3D distance was used [16]: Analysis was performed to compare the accuracy of the AO value from the ground truth to the AO value calculated from the R3D point cloud using FARO Zone 3D.The results of the analysis at each experimental distance tested are displayed in Figures 16-18.
The minimum error, maximum error, mean error, and standard deviation are outlined in Tables 1-3.These tables highlight how the error rates increase as the distance from the impact rig to the wall increases.
It should also be noted the most deviation from the ground truth, regardless of distance, occurred in the height estimation, or z coordinate.The distribution of total errors at 25 cm in the z-axis had a minimum of −0.2 cm, first quartile of 2.2 cm, median of 5.9 cm, third quartile of 5.9 cm, and maximum of 6.8 cm.The distribution at 50 cm in the z-axis had a minimum of −6.9 cm, first quartile of 9.   4).
A summary of data for all six comparisons is provided below.

| DISCUSS ION
The purpose of this research was to determine if R3D had the capabilities of producing high-quality scan data that are sufficient for calculating the AO in FARO ZONE 3D.One of the most common issues encountered with calculating AO with existing methods of BPA (i.e., traditional stringing methods and digital programs) is inaccurate height calculation (z coordinate).This occurs due to the fact that blood does not travel in a linear trajectory.Since the current calculation methods rely on a linear formula for calculating AO, it is unsurprising that the most deviation occurred on the z-axis [1,3,12].This occurs because the calculation of AO and trajectory is based on a linear bloodstain flight path instead of a curved flight path.Blood is a non-Newtonian fluid and has unique composition that affects its density, viscosity, and surface tension [5].As blood travels through the air, it is also influenced by gravity, resistance, and temperature of the surrounding environment.The further the impact rig was moved from the target walls, the less accurate the measurement became, with the z coordinate being F I G U R E 1 6 Errors at individual and 3D errors reported for 25 cm impact patterns.

F I G U R E 17
Errors at individual and 3D error rates reported for 50 cm impact patterns.
the least accurate axes.The large displacement of the z coordinate is well documented in related research, with the current results reflecting past findings.[4,6,27].
When looking at the comparison data between the FARO S350 laser scanner and R3D, it was shown that on average, 91.9% of the points were below 3 mm deviation between scans.At the 95% threshold, the average difference falls below 3.6 mm.These values are absolute differences and a more conservative approach than taking signed differences which would provide a smaller average result.When considering the contribution of error of the R3D point cloud data alone and comparing this to the differences in results of the impact tests, the contribution of error of the R3D data to the overall area of origin analysis is within a few millimeters.When considering the repeatability of each of the scans, there appears to be a greater sensitivity at the lower thresholds.For example, at 1, 2, and 3 mm deviations, the standard deviation was 4.9%, 3.3%, and 1.6%, respectively.
It should be noted that there were some small differences between each of the R3D scans and in some areas, there was some missing data near the baseboard or in other areas.These were not detrimental to the analysis, but simply notable.Increasing the scanning time, positioning the sensor perpendicular to the surfaces, and F I G U R E 1 8 Errors at individual axes and 3D error rates reported at 100 cm impact patterns.

F I G U R E 19
Color ramp applied to Recon-3D scan 1, showing absolute differences when compared to the FARO S350 laser scanner data.
Blue color shows less deviation while red shows greater deviations.
overlapping these areas would have filled the areas with more data.
In addition, the color data of the R3D data appear to provide a more realistic view of what is seen by the human eye.The colors are not washed out and provide slightly better contrast as shown in Figure 6.
The results produced from this study are promising and indicate that R3D is an application with great potential to becoming a powerful and accessible forensic investigation tool.R3D was successful at rendering a complete and useable point cloud through the cloud processing option.The scans produced high-quality point clouds acceptable for capturing small details that are necessary in the field of BPA.Specifically, this study revealed R3D would be highly beneficial to the practice of BPA for stains 50 cm or less from a wall, producing calculated AOs that do not differ significantly from the known coordinates of a blood source.The systematic review published by Home et al. [28] highlighted the large range of accepted error rates that exist with BPA software for trajectory analysis, stating it can be difficult to compare accepted error rates between different software applications.The commonly reported error rates for BPA were considered as they relate to HemoSpat, FARO Scene, and FARO Zone 3D.There is a large amount of data available, and the entirety of the publications was not applicable to review for this study.Further studies should be explored to determine if the large errors are due to R3D or linear trajectories in general.
The studies below were chosen for comparison based on the experimental conditions, software methodology, media, substrates, and distances which most reflect the current study.The parameters and details of the studies that were chosen for comparison are outlined in Tables 5 and 6.The publications were separated based on the maximum distance the impact device was situated from the wall during the experiments, 50 and 100 cm, respectively.Considering the wide range of experimental setup conditions and parameters, it is difficult to draw any definitive comparative conclusions to the relevant research.In order to draw general comparisons, the known values as well as the mean error rates of the current study are listed at the bottom of each table.Most of the error rates recorded F I G U R E 2 0 Histogram for one of the Recon-3D wall scan comparisons.Note that the percentage of points below 1 mm shown as approximately 53%.
photogrammetry, laser scanning, and other 3D technologies have increased in popularity and become accessible and powerful 3D scanning application available to Apple customers with LiDARenabled devices.The app is intended for practical use within law enforcement and forensic professions.The R3D app combines iPhone LiDAR data and video information to allow the user to quickly document scenes or evidence of various sorts.This innovative technology offers a powerful and affordable handheld application capable of capturing high-resolution 3D scans in minutes.R3D provides a simple platform to record and document crime scenes and objects of interest during evidence documentation and recovery.F I G U R E 2 A single 12 × 12 pattern (left image) which shifts to make up the overall pattern (right image).
After naming the scan, inputting AprilTag distance measurements, and choosing the desired resolution, the user can start the scanning process.The length of the scan is chosen by the user, with the application offering a maximum scan time of 10 min.Once the scan has begun, the video information, depth map, and point map appear on the device, providing the user with feedback about the data being collected.To capture the level of detail necessary for good-quality scans, it is recommended to follow good scanning practices.Regardless of the length of the scan, the user should always be moving the device and make sure they are varying the distance between the device and the area of interest.It is also important to change the angle of the device so that the sensor is perpendicular to the surface of interest.It is also useful to capture wide angles of the scene as well as close-ups.If AprilTags are being used, it is important to confirm that the application has detected them, which will be visible as green squares over the targets.Once the user stops the scan, they are given a choice of two different data processing methods.R3D offers on-device processing or cloud processing.Device processing utilizes the devices' operating system to produce a 3D point cloud that can be viewed within the application and is completely offline.Cloud processing sends the data to a cloud-based processing system and notifies the user via email once it is complete.The resulting 3D data are in the form of a point cloud F I G U R E 3 Example of AprilTags as reference targets used on a floor.The distance between AprilTag centers is measured and entered into Recon-3D prior to scanning.F I G U R E 4 Recon-3D scanning parameter options showing the Target Distance field where the value of the distance between target centers is entered.

E 5
Impact rig showing multiple adjustment points as well as the center target for blood.F I G U R E 6 3D-printed square scales, 10 × 10 cm and 20 × 20 cm.blood to use, as well as the best height to release the piston from, a series of tests were conducted with artificial blood from each distance from the wall.These tests concluded that 2 mL of blood struck from a piston drop height of 60 cm was the best combination to create impact patterns with a wide range of stains.A total of eight impacts were created at each distance, resulting in a total of 21 usable impact patterns that were suitable for analysis.Sheep blood was acquired from the Canadian Food Inspection Agency and stored in a 4°C refrigerator.The blood was ethically sourced, pathogen free and contained 1% sodium fluoride as a preservative and anti-clotting agent.The blood was removed from the refrigerator prior to experimentation and allowed to warm to approximately 37°C to mimic normal human physiological conditions.The sample container was carefully agitated to mix blood prior to use, ensuring the mixture contents were homogenous.

F I G U R E 7
captured using the R3D application on a 2021 iPad Pro.To ensure each scan recorded the scene consistently, the same general sequence of scanning movements was performed during each R3D scan.The most important steps throughout the scanning process were to constantly move within the scene, alternating the distance the device is away from the walls, and using different angles.During the first pass of the scan, it was important to capture the plane containing the AprilTags to ensure R3D recognized the targets for accurate scaling (Figure11).To determine the appropriate scanning parameters, a set of four scans were conducted for the first three impacts created at 25 cm and assessed in CloudCompare for F I G U R E 8 2 mL of sheep blood loaded onto the striking surface of the impact rig.F I G U R E 9 Square scale place around at least three areas of clusters for each impact.Experimental setup showing impact rig and AprilTag placement.accuracy and completeness (Figure 12).CloudCompare is a free program that can be downloaded for use on both Windows and Apple operating systems and is used by 3D specialists for quickly and efficiently comparing and examining point cloud data.The software also provides additional tools to modify point clouds, as well as the option to record and export short animations of the point cloud data.The largest variables influencing the R3D scan data revolve around the chosen scan resolution, scanning time, and overall scanning technique.Due to the level of detail needed to analyze bloodstains accurately within a 3D environment, it was determined the resolution must be set at the finest possible setting (1 mm), and a scan time of at least 45 s was necessary.Due to the variable turnaround time of cloud processing (1-3 h), at least two scans were completed at each impact (45 s and 1 min) for redundancy.Once cloud processing was complete, R3D provided an e57 point cloud file that was opened and inspected in CloudCompare prior to analysis in FARO ZONE 3D software.After all the photographs and scan data were collected, the photographs were uploaded to Adobe Photoshop to correct for basic distortion and perspective issues.Before exporting, the photographs of the clusters were cropped to a square the same size as the inner measurements of the square scale in which the cluster was located.Small adjustments were made to the levels of the photographs to correct for exposure and black/white balance.The square scales were very helpful during this step of the data processing, allowing the photoshop workflow to be completed with ease when aligning and scaling the photographs.With completed scans selected for analysis, the e57 point cloud was imported to FARO ZONE 3D.The cloud alignment tools were used to correctly position the point cloud within the 3D space.'Pick point x, y, z' and 'elevate to pick point z' were the tools used to place each point cloud at the same elevation, same coordinate system, and with the same origin point to ensure continuity.This was done by adjusting the translation, yaw, pitch, or roll of the scan.After the point cloud was accurately placed and positioned, the blood spatter plugin, located among the crime power tools within FARO ZONE 3D, was used to import the photographs of the corrected bloodstain photographs with square scales associated with the scan.Once the photo was opened, the program needs to determine how to align and scale the photo with the point cloud in the scene.The 'Align 3D' option was selected for placement, and three points were chosen within the photograph and paired with three matching points on the point cloud.The program is robust, and only small adjustments were necessary after the initial placement of the photograph.The transparency slider was used to compare the placement of the photo on the point cloud with accuracy.This process was repeated for each of the photographs that were captured

F I G U R E 1 0
Photographs were captured using a Nikon D7500 camera with a 1:1 60 mm Nikon Macro Lens.F I G U R E 11Screen capture of on-device processing.The green squares pictured in the center of the AprilTags indicate that Recon-3D has successfully detected the AprilTags.A line was created from the known point to the middle of the estimated AO created in FARO ZONE 3D.This provided a helpful visual cue during the ellipse drawing over the series of photographs and a reference for later calculations.This value was recorded and represented the overall 3D distance from the known impact location to the calculated AO in all three axes, or total 3D error.The individual x, y, z coordinates of the calculated AO were also recorded and compiled into a Microsoft Excel document for later statistical analysis.

F I G U R E 1 2
Assessing an e57 point cloud file produced by Recon-3D in CloudCompare for completeness and accuracy.F I G U R E 1 3 FARO ZONE 3D interface showing the process of overlaying scaled photos to point cloud.Three-point alignment requires the first two points to be generally horizontal (see lower corners of the square photograph), while the third point can be above or below the first two horizontal points.F I G U R E 14 Screen capture from FARO Zone 3D showing the ellipse marking screen.Each ellipse has its own properties and can be adjusted manually in cases where the analyst wishes to refine the fit of the ellipse.Equation (2) represents the Distance formula: The square root of the squared difference of the x values plus the squared difference of the y values plus the squared difference of the z values.
The distribution of total errors at 25 cm in the x-axis had a minimum of −5.3 cm, first quartile of −3.78 cm, median of −2.4 cm, third quartile of −1.25 cm, and maximum of 2.3 cm.The distribution at 50 cm in the x-axis had a minimum of −9.3 cm, first quartile of −5.3 cm, median of −2.9 cm, third quartile of −2.25 cm, and maximum of 4.1 cm.The error distribution at 100 cm in x-axis had a minimum of −22.8 cm, first quartile of −5.75 cm, median of −0.1 cm, third quartile of 3.2 cm, and a maximum of 9.5 cm.The distribution of total errors at 25 cm in the y-axis had a minimum of −4.1 cm, first quartile of −3.6 cm, median of −2.65 cm, third quartile of −1.2 cm, and maximum of 2.4 cm.The distribution at 50 cm in the y-axis had a minimum of −5.6 cm, first quartile of −3.98 cm, F I G U R E 1 5 FARO S350 scan data (left).Recon-3D scan data (right).
median of −1.4 cm, third quartile of −0.025 cm, and maximum of 2.6 cm.The error distribution at 100 cm in y-axis had a minimum of −14.4 cm, first quartile of −6.43 cm, median of −4.5 cm, third quartile of −2.63 cm, and a maximum of 7.3 cm.
).These results are visually represented using a type of color ramp with a histogram scale as shown in Figure 19.The user can choose what the different colors represent, and, in this study, blue colors show differences closer to zero, while red colors show values closer to 10 mm.The color scale progression begins at blue, then green, yellow, and finally red (Figures 19 and 20; Table

able alternative to virtual stringing programs, such as HemoSpat with similar errors when determining the AO of impact patterns [7, 17-19].
There are a few different terrestrial laser scanners which are used primarily in the forensic area for crash and crime scene documentation.These compact instruments are portable with extremely advanced performance capabilities.These scanners use LiDAR technology, which typically uses infrared laser light to facilitate the process of measuring the arrangement and structure of a physical environment.LiDAR-based systems send out light from a laser and record the time it takes for a pulse to return to a sensor on the de-