摘要
In order to solve the problem of inaccurate tracking of the vehicle ahead and the inadequate timeliness of collision pre-warning in a vision-based pre-warning system for reducing traffic safety accidents, a high-accuracy vehicle tracking algorithm based on digital image correlation (DIC) is proposed. With the detected vehicle area as a reference sub-area, the algorithm adopts DIC method to perform sub-pixel level search and matching around reference sub-area, and then implements vehicle tracking. The scaling value of resulting tracked frame can be used to calculate high-accuracy relative speed for enhancing safety assurance. The results of experiments verify that the algorithm proposed can effectively improve tracking accuracy and real-time performance and reduce the dependence on high-resolution lenses.
- 出版日期2018
- 单位合肥工业大学