Object Tracking Based On Color Feature

In realistic life, many meaningful information of vision are included in sports. Detection and tracking of moving object in image sequence is one important area in the domain of digital image processing and computer vision. Its application include fields like robot navigation, vision-based supervision, security surveillance, medical image analysis, industrial automation, video processing etc. Thus, there are great realistic meaning and practical value by researching the detection and tracking of moving objects. This paper realizes the detection and tracking of the moving object, single object automatically tracking have been discussed.The presently common methods on detection and tracking of moving objects are introduced first. For moving object detection, three methods, including'differential between adjacent frames','background subtraction'and'optical flow', are discussed. Through comparison and analysis, the merits,weaknesses and application field of each method are concluded. For moving object tracking, there are three feature matching algorithms:'absolute-balance searching','normalized cross-correlation matching'and'histogram matching'. Two fast matching methods,'sequential similarity detection algorithm'and'pyramid hierarchical searching algorithm' are also introduced. This paper based on OpenCV and Camshift algorithm to realize the self-motion tracking of colorful sequential imaging. This system can detect and track moving object more accurately and less time-consuming. In order to improve the accuracy and enhance the real-time ability of tracking, a modified CamShift (Continuously Adaptive Mean-Shift) algorithm is presented in this paper. we have only calculated color probability distribution of the tracking area, and cleaned the S,V weight which have strong interference to H weight. Experiments show this method has avoided the mistake in the process of tracking and reduced the operation consumedly, it can track the object accurately in a real-time way.The algorithm and methods applied in this research can be applied in fields such as vision-based safety-monitoring, air surveillance, and industrial detection.