Anti-Jamming Tracking Algorithm for Ship Target Based on Correlation Filtering

Abstract

For solving the problem of tracking drift or loss caused by morphological changes and scale scaling in ship target tracking, this paper designs a target tracking algorithm based on multi feature fusion and threshold selection to resist morphological changes and scale transformation interference. In the anti morphological change module, a multi feature weighted fusion method is designed. Through the adaptive weighted fusion of histogram of oriented gradients (HOG), local bninary patterns (LBP) and color names (CN) by the contribution rate of color moment feature recognition, the feature extraction ability of important parts of ship targets is strengthened, and the robustness of the proposed algorithm in the tracking process is improved. In the anti scale transformation interference module, a method with multi resolution target box joint search method and determination of target position by the maximum response peak is designed to solve the problem of low robustness of the tracking box caused by the scaling of ship targets. The experimental results show that the proposed algorithm has better tracking performance, with an accuracy of 93.6% and a success rate of 70.1% on the OTB dataset. This method is superior to other related algorithms.