Assessing agreement among crime scene measurement methods

Abstract

A critical concern with crime scene documentation is the accuracy with which a crime scene can be reconstructed. Here, we discuss the accuracy of eight documentation methods as a function of measurement distance between reference ground targets in an outdoor scene. The relative accuracy of each documentation method was assessed with respect to a widely accepted and well-established standard method for land surveying, Total Station, from which measurements served as "ground truth" or reference data. For the majority of methods, the actual relative difference between measurements when compared to Total Station was small (less than a quarter of an inch). Measurements from FARO LiDAR agreed the most with to those of Total Station, while drone without the use of ground control points (GCPs) agreed the least. GCPs or a reference scale were also found to be important in mitigating increasing imprecision with increasing distance when measuring between two targets ~9-85ft apart via drone and orthomosaic methods. Additionally, there were no statistical differences in the use of 2D (horizontal) or 3D (slope) measurement configurations for the Total Station. Overall, linear regression of difference plots did not reveal meaningful correlation between increasing distance measured and the error of a method when compared to Total Station. As more measurement methods become available, and the need for training and validating new tools become a necessity, these results point to the importance of establishing a ground truth or known distance range on which crime scene measurement methods can be validated.