Cross-lens multi-target tracking method and device based on space-time constraint

Abstract

The present invention discloses a cross-lens multi-target tracking method and a device based on space-time constraint. The method includes: performing image preprocessing on different color spaces so that the images are consistent in color temperature and hue to obtain the photographing information of multiple photographing devices; establishing a corresponding relationship between 2D points through the projection matrix of the photographing devices to obtain geometric information among the multiple photographing devices, wherein the projection matrix is a projection matrix about the 3D world; and performing human body characteristics matching among the multiple cameras according to the photographing information and the geometric information to take advantage of the apparent and temporal characteristics of the tracking target to obtain the pictures and the real time tracking result of each photographing device. By combining the current multi-target tracking algorithm with the multi-camera processing method and using the attitude and posture matrix of the camera network, the method can achieve multi-target tracking based on multiple cameras. This improves the robustness of target tracking but reduces the tracking error, therefore, increasing the tracking accuracy.

Description

technical field [0001] The invention relates to the technical field of visual target tracking in computer image processing, in particular to a cross-shot multi-target tracking method and device based on time-space constraints. Background technique [0002] Video object tracking refers to given the initial position of the object in the video, and then outputs the position of the object at every moment in the video. Object tracking is an important problem in computer vision and is usually the first step in video analysis processing. Therefore, a large number of scholars are engaged in the research of object tracking, and many effective object tracking algorithms have been proposed. In some monitoring scenarios, multiple objects need to be tracked simultaneously in a complex scene. Mutual occlusion between multiple objects increases the difficulty of object tracking, which often occurs in pedestrian tracking. When a large group of people appear in the image of the camera equ...

AI Technical Summary

Problems solved by technology

Among them, the repeated coverage area in the cross-camera tracking problem and the processing method of the non-covered area have been discussed in many articles. With the needs of security and pedestrian data analysis, multi-target tracking based on multi-camera is very meaningful. But at the same time, due to the complexity of its problems, this work is also very challenging

Recently, some scholars have proposed a variety of ways to use information from multiple cameras to improve the robustness of object tracking, but they ignore geometric constraints and other issues, violate geometric assumptions, and require more complex methods to solve the resulting errors.

Images