Moving Object Detection And Tracking For Volleyball Video

Sport videos are becoming more and more popular. Therefore, it has highly practical significance and commercial value to analysis sport videos. We researched on the moving object detection, prediction tracking and trajectory in volleyball video sequences captured by a fixed camera. The objects include players and volleyball.This paper, volleyball video sequences are pre-processed, including image denoising, image enhancement and morphologic transform, etc. By those operations, the moving objects in image are easily identified, the image quality and the accuracy of the detection algorithm is improved. Take into account both quality and real-time, this paper improved median filter algorithm. For moving objects detection, a background subtraction algorithm based on adaptive Gaussian mixture model is proposed in order to extract moving regions. The OTSU algorithm is researched in order to adapt to the changes in the background images; In order to accelerate model updating rate, a novel mechanism is the combination of expected sufficient statistics and L-recent window.For moving objects tracking, firstly, the ID and location information of the objects blob are received through image segmentation. Secondly, tracking algorithm is a combination of centroid tracking algorithm and Kalman filter. The algorithm can provide players and volleyball object tracking. In the process of tracking, Kalman filter is imposed to forecast the motion information to improve the tracking precision.On the basis of the moving object tracking, this paper takes path generating module to describe players tracks and preserve motion information; Based on the principle of double camera measure to reconstruct 3D information by two images taken from different angles at the same time, 3D information is used to match the volleyball trajectory.Taking advantage of the information of the moving objects in volleyball video, this paper consists of two parts including theoretic and pragmatic researches on the theory of moving object detection and tracking and has obtained corresponding results. Experimental results indicates that the rate of accuracy of this algorithm higher than 92%, the algorithm can realize the real-time and stable while ensuring the accuracy, providing necessary basis for accurately video analysis.