Coded Target Based Traffic Accident Vision Swift Reconnaissance System

Aimed at the defects of current traffic accident rapid investigation methods, a traffic accident vision swift reconnaissance system based on coded target is proposed. Make use of the uniqueness and feature station´s geometric constraint of coded target, and we can realize automatic recognition and match of coded target and precise space scale constraint. The coded targets distributed in the accident scene are imaged by the CCD camera in several locations and orientations, and the 3D coordinates of space measured feature points are automatically solved through the initial external location and orientation of camera based on co-angular constraint pose estimate algorithm and full-scale bundle adjustment algorithm based on collinear constraint. Therefore the 3D information of the traffic accident scene is swiftly acquired and the traffic accident scene is also reconstructed. The experiment results demonstrate that the proposed method is feasible and flexible.