YOLOX-based end-to-end multi-target tracking method and system

The invention provides an end-to-end multi-target tracking method and system based on YOLOX, and belongs to the field of deep learning target detection and tracking. According to the YOLOX-based end-to-end multi-target tracking method, aiming at the situation that an existing target tracking method needs to use two steps of detection and tracking, an end-to-end target tracking method realized by using a YOLOX-based target detection algorithm is provided, and the multi-target tracking method can be realized by only needing one YOLOX network. According to the algorithm, firstly, the input end of a YOLOX target detection algorithm is improved to adapt to target tracking of a video stream; and aiming at the problem of low speed caused by detection and tracking separation of the existing target tracking method, the appearance characteristics and position prediction of the target are integrated into the target detection network in the YOLOX network, so that the end-to-end target tracking method is realized, and the