Moving target tracking method based on kcf and human brain memory mechanism

Abstract

The invention discloses a moving target tracking method based on KCF and human brain memory mechanism. The visual information processing cognitive model based on the human brain memory mechanism is introduced into the target template update process of the KCF algorithm, and the memory-based Template Updating (MTU) method forms a new model update strategy, so that each Each template has to be transmitted and processed through the three spaces of short-term memory, short-term memory and long-term memory. During the target tracking process, the target template is updated according to different update strategies according to the matching degree between the current frame target template and the target template in the memory space. By effectively memorizing the previous scene, the method can continue to track robustly when the current target has a sudden change in attitude, reappears after a short disappearance, or occlusion occurs.

Description

Technical field[0001]The present invention relates to a tracking method of moving target in an image sequence, specifically, a moving target tracking method based on a nuclear correlation filtering algorithm (KCF, Kernelized Correlation Filter) and a human brain memory mechanism, belonging to computer vision technology.Background technique[0002]Sports Target Tracking is an important research direction in the field of computer visual, widely used in intelligent security, visual monitoring, human-machine interaction. In recent years, although the target tracking technique has made great progress, it is still difficult to achieve accurate target tracking under complex conditions such as light changes, target geometric deformation, target occlusion, rapid movement.[0003]At present, the target tracking method based on the discrimination has become mainstream, including: Structure Output Tracking With Kernel, Struck Tracking Method, Tracking - Learning - Detection, TLD) Tracking Method, V...

AI Technical Summary

Problems solved by technology

[0005] On the other hand, although the KCF algorithm has advantages, it also has certain limitations: first, it cannot deal with the problem of complete occlusion of the target; second, it lacks an adaptable model update strategy

Therefore, when the moving target has a sudden change in attitude, the target reappears after a short disappearance, or the target is occluded, the tracking accuracy of the KCF algorithm will drop significantly, or even fail to track.

Images