Assessing Vehicle Profiling Accuracy of Handheld LiDAR Compared to Terrestrial Laser Scanning for Crash Scene Reconstruction

Abstract

Forensic crash investigation often requires developing detailed profiles showing the location and extent of vehicle damage to identify impact areas, impact direction, deformation, and estimated vehicle speeds at impact. Traditional damage profiling techniques require extended and comprehensive setups for mapping and measurement that are quite labor- and time-intensive. Due to the time involved, this damage profiling is usually done in a remote holding area after the crash scene is cleared. Light detection and ranging (LiDAR) scanning technology in consumer handheld electronic devices, such as smartphones and tablets, holds significant potential for conducting this damage profile mapping in just a few minutes, allowing the mapping to be conducted at the scene before the vehicle(s) are moved. However, there is limited research and even scarcer published literature on field procedures and/or accuracy for these emerging smartphones and tablets with LiDAR. This paper proposes a methodology and subsequent measurement accuracy comparisons for survey-grade terrestrial laser scanning (TLS) and handheld alternatives. The maximum root mean square error (RMSE) obtained for profile distance between handheld (iPad) and survey-grade TLS LiDAR scans for a damaged vehicle was observed to be 3 cm, a level of accuracy that is likely sufficient and acceptable for most forensic studies.

Reconstructing 3D model of accident scene using drone image processing

“…Amin et al [24] stated that the accuracy for the POI technique was 0.059 m and the waypoint technique was 0.043 m. Desai et al [34] mentioned that the scale errors of 3D reconstructed accident scene using the waypoint technique was about 0.039 m. Almeshal et al [35] achieved the accuracy of 3D accident scenes was about 0.047 m. This study has produced 3D results of accident scenes with an accuracy of 0.047 m for the circular technique and 0.06 m for the double grid technique. Therefore, the results achieved in this study were nearly similar to the previous studies.…”

Automating building element detection for deconstruction planning and material reuse: A case study

“…The scaling was checked using manual spot measurements of the distance between the flat faces of the columns to ensure that it could be properly compared to the ground truth. Although generally measuring true distances, the Apple Lidar system has an observed per-point mean error of 1-2 cm [37,76], and total error at the building scale of up to 10 cm [77]. As this final discrepancy is due to non-linear errors during the mobile trajectory reconstruction, there was no attempt to correct it using scaling.…”

Assessment of Indiana Unmanned Aerial System Crash Scene Mapping Program

“…Even though both measuring tapes and electronic total stations can provide detailed two-dimensional (2D) measurements, three-dimensional (3D) documentation of traffic crashes cannot be obtained from traditional 2D measuring tools. The emergence of terrestrial laser scanning (TLS) has enabled investigators to observe the type of entity damage and analyze contributing factors such as vehicle dynamics using a processed 3D digital model [21][22][23]. Given the large amount of data gathered, the speed of data acquisition was considered fast using TLS.…”