Vehicle Movement Information Reconstruction Based on Video Images and DLT Theory

Xueyuan Han,Xianlong Jin,Xiaoyun Zhang,Xinqiang Miao
DOI: https://doi.org/10.3969/j.issn.1000-680X.2012.12.017
2012-01-01
Abstract:By applying direct linear transformation theory in photogrammetry and based on the marked points information in accident site, the 2D direct linear transformation coefficients of orthographic projection transformation are solved out. By frame decoding of accident surveillance video, a group of pictures are obtained, and then a series of contact points between wheel and road are marked on decoded picture group and an orthographic projection is conducted on the marked points on wheel. Based on these marked points the trajectory curve is fitted, and the length of trajectory is calculated by integration. Combined with frame rate information of video the speed of accident vehicle is finally obtained. The method is applied to a real traffic accident case with the data of driving trajectory length and vehicle speed calculated.
What problem does this paper attempt to address?