Cross-video target tracking method and system, electronic equipment and storage medium

Abstract

The invention relates to a cross-video target tracking method and system, electronic equipment and a storage medium. The method comprises the following steps: determining a to-be-tracked video image and an initial target template image; inputting the video image and the initial target template image into a trained deep twin network tracking model, and outputting a target prediction image of the tracking target in the video image through the deep twin network tracking model; mapping the video image into a unified geographic coordinate space by using the trained geographic mapping model to obtain a global geographic coordinate of the video image; and on the basis of the mapped video images, calculating an overlapping region between every two video images through a polygon cutting method, judging whether a target prediction image exists in the overlapping region or not, and performing cross-video target tracking in the video image corresponding to the overlapping region with the target prediction image by using a tracking target handover algorithm. According to the invention, the target can be continuously and stably tracked for a long time, the real-time performance of tracking is ensured, and the tracking precision is high.

Description

technical field [0001] The present application belongs to the technical field of visual tracking target tracking, and in particular relates to a cross-video target tracking method, system, electronic device and storage medium. Background technique [0002] With the development of science and technology, people's quality of life continues to improve, and the demand for safety is becoming more and more urgent. The video security monitoring industry has developed rapidly, and various applications have also put forward different monitoring technology requirements. Video security monitoring has gradually moved from digitalization and gridization to intelligence. Intelligent monitoring systems are often used in social security places, large entertainment places and various road traffic places. In intelligent monitoring systems, visual tracking target tracking is a basic requirement, and it is also the basic task of subsequent advanced visual processing such as gesture recognition ...

AI Technical Summary

Problems solved by technology

However, in the scene of multi-camera collaborative cross-video tracking, due to the scene transformation caused by multi-video switching, problems such as scale change, appearance change, illumination change, occlusion and disappearance of the tracking target will occur, resulting in unstable tracking of the tracking target and difficult to achieve. Cross Video Tracking

Secondly, the tracking duration of cross-video tracking targets is longer, and it is easier to produce changes in scale and appearance than short-term tracking targets. The tracking model will produce tracking drift due to the accumulation of tracking errors, resulting in tracking failure.

In addition, cross-video tracking involves joint analysis of multiple videos, which requires high real-time reasoning speed for tracking algorithms, and complex tracking algorithm models cannot be directly applied

Images