A Correlation Filter Updating Device Considering Scene Classification in Target Tracking

Abstract

The invention discloses a correlation filter updating device considering scene classification in target tracking, comprising: a target block module, a parameter acquisition module, a target scene judgment module and an adaptive learning update module. In the actual target tracking process, it will be affected by many realistic conditions, such as the deformation of the target itself or the target being occluded, especially for long-term and large-area occlusions. Due to the limited understanding of these scenes by existing correlation filters, It will greatly degrade the performance of the tracker. In order to improve the robustness of tracking, the present invention proposes an update device for adaptive learning, which adaptively updates the correlation filter for different scenes where the target is located, and solves the problem of different scenes such as long-term large-area occlusion and changes in the appearance of the target. Track performance impact.

Description

technical field [0001] The invention belongs to the technical field of target tracking, and in particular relates to a correlation filter updating device considering scene classification in target tracking. Background technique [0002] Moving target tracking is a very important part of the application in the field of computer vision, and can be applied to various fields such as video surveillance, human-computer interaction, visual navigation, and military guidance. Moving target tracking uses a tracking algorithm to analyze and process video image frames to obtain the position of the target. However, in the actual tracking process, it will be affected by many realistic conditions, so that the real-time performance of the tracking algorithm cannot meet the requirements. [0003] In response to this problem, scholars have applied the idea of ​​correlation filtering to target tracking, which has improved the speed of target tracking. After that, many improved algorithms base...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the deficiencies of current target tracking methods based on correlation filtering, the purpose of the present invention is to provide a correlation filter update device that considers scene classification in target tracking. The scene adaptively updates the correlation filter to improve the accuracy of tracking, and uses discrete Fourier diagonalization to reduce the complexity of storage and calculation

Images