The Moving Target Detection And Tracking On The Video Of Robocup Martial Arena Contest

At present, the moving target detection and tracking is a very valuable direction in the area of visual identification research. It has wide application in civilian? military and other aspects of life. Such as in the guidance of aviation and military aircraft, UAV autopilot, the weapon aiming system based on video, traffic monitoring, and the analysis of the medical biological tissue. Therefore, it is essential to accurately capture the moving objects in complex scenes.In this paper, we mainly study the tracking algorithm of moving objects in the case of camera fixed, and some new improved methods are forward in this article, which based on the tracking algorithm of Kalman filter. First of all, this paper describes four common target detection methods, and analyzes the advantages and disadvantages of these algorithms, as well as the target detection algorithm that used in this paper. Secondly, the current several mature motion objects tracking algorithms are introduced, such as tracking algorithm based on regional matching, tracking algorithm based on active contour tracing, tracking algorithm based on motion characteristics, and tracking algorithm based on neural network. Focuses on analyzing and illustrating the classical algorithm of Kalman filter, the basic principle of the algorithm is described, the state equation and observation equation of Kalman filter are listed, the implementation process of the algorithm is analyzed, and the shortcomings of the algorithm are summarized. And based on the analysis of classical algorithm, proposing an improved Kalman filter algorithm, which greatly improves the real-time performance of the algorithm, and it is effective to track the moving objects.Finally, through the two videos, a video shooting at the scene of the match, another is a recorded video, under the environment of MATLAB simulation software, by the tracking of moving objects in the two video, it is proved that the algorithm is real-time and effective, what's more, compared with the tracking results of the traditional Kalman algorithm. It can be seen from the tracking results that the improved Kalman tracking algorithm is significantly better than the traditional tracking algorithm. At the same time, it still has a great tracking effect when the moving objects in great changes of the speed. It is proved that the algorithm that presented in this paper has better robustness and tracking quality, optimization and improvement for the previous Kalman tracking algorithm successfully.