关键词: 目标跟踪, 粒子滤波, 相关滤波, 自适应空间正则项, 畸变感知

Abstract: A distortion-aware correlation filter object tracking algorithm is proposed to address the problem that the existing correlation filters have insufficient ability to deal with target distortion and the filter model updating error accumulation easily leads to tracking failure. The particle filter is used to construct spatial reference weight that reinforces the target information while suppressing the background interference information; Meanwhile, to optimize the algorithm and reduce computational complexity, the alternating direction multiplier method is used to solve the objective optimal function value with fewer iterations; To further enhance the discriminative capability of the filter, a target distortion-aware strategy is designed to use particle filter for selective target re-detection when target tracking and localization accuracy is poor. Depending on the extent of distortion of the tracking target at any given time, the filter model is adaptively updated. Comparing experiments with various representative correlation filters on the OTB50 and OTB100 datasets, the experimental results show that the tracking success rate and accuracy of the distortion-aware correlation filter object tracking algorithm are the best, it has strong robustness in the face of targets distorted by multiple interference factors in the actual scene.

Key words: object tracking, correlation filter, particle filter, spatio-temporal regularization, distortion-aware