An unmanned aerial vehicle (UAV) is introduced to monitor the traffic accident site for traffic accident scene reconstruction. First, the UAV is used to obtain the sequential images of the traffic accident site. Based on these images, the structure-from-motion algorithm is implemented to present a 3D sparse point cloud model, and the multi-view stereo algorithm is conducted to generate a 3D dense point cloud model. Next, the meshing and texturing operations proceed to optimize the point cloud model, and the digital surface model and digital elevation model of accident scene are provided. An investigation experiment is conducted for accident scene reconstruction. In addition, two evaluation indicators are adopted to assess the quality of the accident scene reconstruction. Six referential objects are selected to measure the reconstruction accuracy. The results of the experiment demonstrate that the proposed method is appropriate for accident investigation and reconstruction.
展开▼