Research On Object Detection And Tracking Algorithms Based On Video Image

With the rapid development of science and technology,computer vision has become a hot topic,and the target detection and tracking technology is an important branch of computer vision.Target detection and tracking are achieved mainly through the correlation algorithm,to detect objects of interest in video sequences,and then determine the position of the object.At the same time,the shape,speed and trajectory of object are described.At present,the target detection and tracking algorithm has been widely used in our lives.However,due to a variety of factors such as the impact of the background environment,the quality of equipment and illumination changes,almost all of the target detection and tracking algorithm has some limitations so that the target detection and tracking effect is very poor in special circumstances,even unable to track the target.Therefore,the design of a target detection and tracking algorithm with good stability,low complexity,high accuracy and adaptability is still an important and difficult research direction.This paper analyzes and summarizes the research status in overseas and domestic,and on this basis,a single target detection and tracking algorithm for video image is studied.The traditional target detection and tracking algorithm has a good result on realtime,tracking,but when the target scale changes in the process of movement,the tracking result is poor.At this time,the traditional approach is not only sensitive to the size of the target,but also lack of real-time update of the template.To improve the shortcomings and deficiencies of target detection and tracking algorithm and get a better tracking effect,this paper points out the influence of control of the distance weighted factor on target detection and tracking algorithm and improved the traditional Mean-Shift algorithm based on video image by extracting analysis of edge weighted histogram and center weighted histogram of target and template.Then comparing the similarity to reflect the changes of target size effectively,and as a basis to realize the adaptive bandwidth of kernel function.