Intelligent Vehicle Infrastructure Cooperative Systems is a key research topic in the field of Intelligent Transportation Systems , and traffic object perception based on cameras is one of the foundations. Due to the development of computer vision, single-camera traffic object tracking has implemented some advanced methods, but cross-camera traffic object tracking is still inadequate in identity matching, especially cross-camera vehicle tracking, because of more similar appearance. With the background of multi-cameras, we take DeepSORT algorithm as the basic framework and propose a vehicle identity matching algorithm based on the re-identification features with confidence determined by spatio-temporal information. The proposed method has been testified on benchmark dataset of traffic video, achieving great performance and verifying its validity. Finally, our work further discusses the advantage and disadvantage of our cross-camera vehicle tracking algorithm based on joint target matching of vehicle features and spatio-temporal information, putting forward the future improvement direction of the algorithm.

CICTP