The Design Of Judgment System Based On Computational Vision

This paper has an analysis based on the current situation in which the image processing technology is used in the judgment system of track-and-field. To meet the requirement, a new kind of judgment system based on computational vision is introduced, when sports meetings are held by universities or middle schools. The new system can solve the problems, such as the waste of labor, lower precision, complicated operation and higher cost. This paper emphasizes on the design of timing program and the image processing algorithm when the athletes pass by.The most critical point when choosing the system hardware are the applicability and the cost. Under such circumstances, a normal camera, video frequency gathering card and PC are included, which had already owned by meeting sponsors. So it has higher cost performance.The system software platform is Borland C++ Builder and Visual C++. In order to guarantee the timing accuracy, the method of high-resolution counter is adopted. The timing precision can reach a millisecond level, which is completely in line with the requirement of system design.Background differencing is used as the image processing algorithm when the athletes pass by which is widely used in the image identification. Meanwhile, the adapted background image method of withdrawing is improved as the combination of mathematical morphology and characteristic point registration so that the robustness is enhanced. Finally, a well-prepared and human natural judgment system is introduced considering by the aspect of user.The computational vision, which the judgment system designed by this paper based on, has a high precision in both timing and the athlete's recognition. It also has other advantages such as low cost, convenience with system's maintaining, applicability in extensive range and good combinability. There is very good application value to put forward to.