THREE-DIMENSIONAL CRUSH SCANNING METHODS FOR RECONSTRUCTION OF VEHICLE COLLISION ACCIDENTS

摘要

This paper presents efficient 3-D collision crush modelling methods utilizing photogrammetry and 3-D scanner for reconstruction analysis of vehicle collision accident. Photogrammetry for 3-D collision crush modelling was executed in two methods: Indirect photogrammetry utilizing small amount of preexisting photographic data, and direct photogrammetry utilizing large amount of photographic data collected for the purpose of 3-D modelling. With the 3-D scanner, it was possible to create a more precise 3-D crush model than the photogrammetry method. Using these 3-D crush models constructed with the mentioned methods, several applications in accident reconstruction analysis are presented, based on Construction methods and characteristics of the resulting crush model. Indirect photogrammetry produced a 3-D model capable of providing approximate shape of crush with only 2 similar to 3 photos of the damaged vehicle, and made it possible to measure the crush volume for estimating the collision velocity. With direct photogrammetry or 3-D scanner, moving and rotating elaborately built 3-D vehicle deformation models makes it possible to precisely reconstruct the actual collision configuration. Additionally, by elaborately modelling the vehicle and its fragments and then putting them together, it is possible to construct a post-collision model of the whole vehicle.