Moving Target Tracking Method Based on Improved Camshift

Abstract

Aiming at the problem that target occlusion and other disturbances in complex background will reduce the tracking accuracy of moving target, and even lead to tracking failure, this paper proposes a moving target tracking algorithm based on the improved Camshift algorithm. Firstly, Gaussian background is used to model the foreground image to improve the backprojection image, and then the interference of backprojection is removed to improve the tracking effect in complex background conditions. Secondly, Kalman filtering is utilized to predict the trajectory, which further improves the tracking accuracy of Camshift algorithm in occlusion condition. A lot of experiments are processed, and the results show that the proposed algorithm could effectively improve the tracking accuracy and meet the real-time requirements.