Moving Target Tracking Based on the Improved Diamond Search Algorithm and Gaussian Scale-Space

摘要

In order to realize an efficient and robust moving target tracking under various scenes, a new template matching algorithm is presented. Firstly, for higher matching efficiency, the improved diamond search (IDS) algorithm was proposed, which increased the matching speed on the basic of ensuring the matching performance. Secondly, the template matching in Gaussian scale-space was put forward to track the object with different scales. Finally, two means were adapted to solve the template drift, one is color histogram, and the other is the modified template update algorithm based on the template buffer. Experiments under complex scenes showed that the proposed tracking algorithm can adapt to such scenes as target' scale change, shape deformation, occlusion and so on, achieving efficient and effective target tracking.