Comparative Analysis Of Object Tracking Algorithm Based On Computer Vision

Tracking research on moving targets has always been a hot topic in the research of image processing.However,it is easy to lose the tracking target because of the interference of occlusion,rotation and light change of moving target in the tracking process.Therefore,how to effectively describe or express the appearance space of interest target in video is a very important part of the tracking process.We firstly increase judgment mechanism in the process of the algorithm of tracking,when the target is lost when introducing SIFT descriptor kernel correlation filter tracking algorithm framework(KCF),and thus improve the performance of the algorithm can track.The experimental results show that the improved method has better tracking effect than the original method.In order to facilitate the comparison analysis between the algorithms,we also designed and implemented a multi-algorithm analysis platform.In this platform,we included some classic tracking algorithms.On the platform,the corresponding interface is designed to facilitate the addition of follow-up tracking algorithm,so that the added algorithm can be compatible with different input forms and performance evaluation.In short,our research puts forward a new idea for the target tracking algorithm on the one hand,and provides a convenient system platform for comparison of different algorithms on the other hand.