Optical flow tracking method and device

Abstract

The invention provides an optical flow tracking method and device. The method includes: generating a random structure according to an object to be tracked on an image; tracking all structure blocks in the random structure by an optical flow tracking algorithm to obtain location offset of the random structure; estimating a target location of the object to be tracked according to the location offset of the random structure. The random structure comprises at least two structure blocks having fixed relative locations. Each structure block comprises at least one feature point of the object to be tracked. The optical flow tracking method and device has the advantages that the feature points of the object to be tracked on the image are acquired, the random structures are generated from the feature points, each random structure is tracked so as to obtain location offset of the random structure, the location of the object to be tracked is estimated accordingly, tracking errors are reduced, and tracking precision is improved effectively.

Description

technical field [0001] The present invention relates to the field of information technology, in particular to an optical flow tracking method and device. Background technique [0002] In the field of video surveillance, it is generally used to detect and track the movement of faces, pedestrians or vehicles through target detection and tracking technology. In the field of video compression and robot navigation, it is also necessary to analyze the movement information of the target. Therefore, the application scenarios of motion analysis are rich and of great value. Motion analysis is to obtain the position offset of the object, so as to track the object. At present, the optical flow method is often used for motion analysis, and the motion direction and motion speed of image points can be determined through optical flow calculation. It requires that the time interval between adjacent images is very small, and no significant change occurs between two adjacent images. [0003...

AI Technical Summary

Problems solved by technology

[0004] However, since different tracking points are easily ambiguous, tracking errors occur, resulting in insufficient tracking accuracy

Images