激光与光电子学进展, 2020, 57 (18): 181023, 网络出版: 2020-09-02
基于归一化互相关匹配算法和Kalman预测器的目标跟踪 下载: 762次
Target Tracking Based on Normalized Cross-Correlation Matching Algorithm and Kalman Predictor
图像处理 目标跟踪 三帧差分法 Bhattacharyya距离 归一化互相关匹配算法 Kalman预测器 image processing target tracking three-frame difference method Bhattacharyya distance normalized cross-correlation matching algorithm Kalman predictor
摘要
针对基于模板匹配的目标跟踪算法在目标被遮挡时容易出现跟踪丢失的问题,提出一种改进的结合遮挡判断和Kalman预测器的模板匹配算法。首先使用三帧差分法提取运动目标并计算运动目标区域。然后针对目标是否被遮挡引入Bhattacharyya距离进行判断,当Bhattacharyya距离小于设定的阈值,表明目标没有被遮挡,则使用归一化互相关(NCC)匹配算法对目标进行稳定跟踪,反之则利用Kalman预测器对被遮挡目标的位置和大小进行预测。实验结果表明,所提算法在静态背景下、目标发生遮挡时的跟踪成功率达到71.43%,比单一NCC匹配算法提高了21.43个百分点。
Abstract
The target tracking algorithm based on template matching can easily lose track of its target when the target is occluded. To resolve this problem, an improved template matching algorithm combined with occlusion judgment and the Kalman predictor is proposed in this study. Initially, the three-frame difference method is adopted to extract the moving object and estimate the moving target area. Then, the Bhattacharyya distance is used to evaluate whether the target is occluded. The target is not occluded if the Bhattacharyya distance is less than the setting threshold, then, the normalized cross-correlation (NCC) matching algorithm is used to track the target in a stable manner. However, if the target is occluded, the location and size of the occluded target are predicted using the Kalman predictor. The experimental results indicate that the tracking success rate of the proposed algorithm is 71.43% when the target is occluded in a static background, which is 21.43 percentage greater than that of the single NCC matching algorithm.
马永杰, 龚影, 陈敏. 基于归一化互相关匹配算法和Kalman预测器的目标跟踪[J]. 激光与光电子学进展, 2020, 57(18): 181023. Yongjie Ma, Ying Gong, Min Chen. Target Tracking Based on Normalized Cross-Correlation Matching Algorithm and Kalman Predictor[J]. Laser & Optoelectronics Progress, 2020, 57(18): 181023.