Based On Background Subtraction Basketball Shooting Automatic Identification System Design

In recent years, with the rapid development of computer vision technology, the moving target detection technology, a significant branch of computer vision, has attracted much attention. Currently, moving target detection technology is widely applied in such fields as military defense, security monitoring, medical examination and intelligent transportation. The technology requires the moving target in video image sequence be separated from background accurately and quickly, it is fundamental to subsequent target recognition, location and behavioral analysis.For the elimination of the drawbacks in fixed-point shooting test system, such as inconvenient installation, complicated making, leak detection and false detection, the paper proposes a method to test shooting by using moving target detection technology. Firstly, the common optical flow method, inter-frame difference method and background subtraction method are analyzed and compared. As the background subtraction has not only the real-time performance, accuracy, but shows the excellent robustness when testing variable speed and non-rigid objects, it is selected as the shooting detection algorithm in this subject research.Then, background subtraction and correlative image processing methods in the process of testing are studied in depth, and background modeling approach, threshold selection method, image denoising method and mathematical morphology are introduced in detail. Through the in-depth research and analysis of the above methods, the shooting detecting algorithm based on background subtraction is implemented by combining background subtraction with detecting characteristics. Considering only the basketball position and status feature in basket part be judged when shooting, the paper uses region segmentation method to segment non-detection region automatically, thus the system detecting speed is further accelerated, at the same time, the image denoising is conducted and leads to the accuracy of basketball position and feature extraction.Finally, after the shooting detecting algorithm is determined, MATLAB is used to design and develop a shooting automatic identification system which is based on background subtraction. The specific design scheme, software and hardware environment setup and user interface design are elaborated, and a detailed introduction is given to the function of each module. Furthermore, by detecting the recorded video, comprehensive evaluation is employed to evaluate the system performance from the aspects of accuracy, real-time performance and stability.