An Adaptive Method for Dynamic Background Compensation Based on SIFT

To address the issue of poor object detection and tracking performance in dynamic scenes, a low-complexity adaptive dynamic background compensation algorithm is proposed to improve the real-time and accuracy of object detection and tracking. The algorithm uses the SIFT feature point matching algorithm to assist in calculating the background motion parameters and introduces an adaptive coefficient with feedback mechanism to reduce the matching calculation range of background motion parameters, thereby achieving dynamic background compensation. While improving the accuracy of background compensation, the algorithm also speeds up the computation to some extent. Experimental results show that this algorithm can improve the accuracy of detecting and tracking moving objects in video sequences with complex dynamic backgrounds while occupying only a small amount of system resources.