The present work aim to verify the effectiveness of modern scanning technologies in the process of reconstruction of vehicle accidents. In classic approaches, deformations of vehicles are mainly identified by visual inspection and/or manual metric measurements; accident scenes are reconstructed point-by-point, using measuring wheel or, at best, laser and photogrammetric instrumentation.
These techniques allow acquiring a very limited amount of data and with a considerable expenditure of time. Modern techniques of computer vision, such as structure from motion (SfM) and multi-view stereo (MVS), and techniques of laser scanning, allow to obtain a dense 3D point cloud of a scene of interest with semi-automated procedures. In the present work, several tests have been performed in order to evaluate the applicability of these technologies to the detection of the shape of a vehicle and to the reconstruction of a scene of an accident. The performed tests showed how new technologies can greatly improve the quality of a reconstruction, producing reliable and complete information that can be easily integrated into specialized software for vehicle accident reconstruction.