Research Of Multi-target Moving Detection And Tracking Method Used For Surface Surveillance On The Airport

The rapid development of the aviation industry make more and more people go into the research of airport surface surveillance technologies. Although existing surveillance technology can meet security needs, but these technologies require more money and higher airport conditions. The purpose of this paper is according to the computer vision technology,using a CCD camera to position moving targets on the airport surface. This noncooperative surveillance technology has high positioning accuracy, low cost and high value of research.Detecting and tracking moving multi-targets algorithms in a sequence of images of the video are the research emphasis. The main task is to simultaneously track multiple moving targets in the video which cameras captured and output their pixel coordinates in real-time.According to the binocular stereo vision system model, pixel coordinates can be transformed to geodetic coordinates, achieving the purpose of tracking.The main research work is as follows:1. In order to detect the target easily, a sequence of video images were preprocessed.After that, several methods for detecting moving targets are briefly analyzed. According to their applicability, an appropriate method is selected and then realized through Matlab platform, which has achieved target detection on the airport surface. Morphological operations are introduced to make the detection results more clear and complete. For multi-target tracking, there are more connected regions in the detection results. Through the connectivity of the pixel, a method to mark the order of targets is used.2. A method of tracking moving objects based on Camshift is analyzed. It is simulated in the MATLAB environment and two videos are chosen to test its effectiveness. The result shows that Cam shift algorithm doesn’t work well when the color of targets and background is similar. In addition, when the target moves too fast, the algorithm is difficult to quickly capture the target, the tracking results are seriously shifted.3. A method of tracking moving objects based on Kalman is analyzed. It is simulated in the MATLAB environment and two same videos are performed. Kalman algorithm can accurately track multiple targets simultaneously, and output their pixel coordinates in real-time. For target who moves fast, such as a taking off aircraft, its tracking performance is also very good.Finally,a summary of this article is shown. The significance of this project is illustrated once again. According to the current shortcomings, some possible improvements are suggested, and the direction of the follow-up study is indicated.