Methods for detecting manipulations in 3D scan data

Abstract While interest in using 3D scanning technology for crime scene investigation (CSI) has grown in recent years, a number of barriers still remain that prevent its wide adoption in the criminal justice system. One such barrier comes from the lack of tools that can validate a 3D scan and verify that it has not been manipulated. While a great deal of research has gone into the detection of manipulations for 2D images, the detection of manipulations for 3D scenes has yet to be fully realized. This paper introduces a series of techniques to detect if a 3D point cloud generated from a LiDAR scan has been subsequently manipulated. These techniques exploit fundamental structures inherent in the collection and storage of these types of data. While the proposed techniques are able to detect a number of different types of manipulations, their limitations are also discussed. The goal of this work is to provide a foundation for the creation of a validation toolkit that can ensure 3D scan data is valid and unaltered.