In order to improve the speed and accuracy of target tracking for missile-borne image, a method for detection, tracking and segmentation based on IoU correlation is proposed. By designing the anchor boxes on the multi-scale prediction branch, an enhanced YOLOv3 detection method is proposed, which improved the target detection capability. Combined with the mask branch, the tracking method based on Siammask achieved multi-target tracking and segmentation. The experimental results show that compared with SiameseNet and KCF, the tracking accuracy of the algorithm is improved by 19.5% and 24.5% respectively, and the running speed is up to 40fps, which meets the accuracy and real-time requirements of target tracking.
