Fusion Image Spatial Information Of Urban Low Altitude Target Tracking Method Research

Moving target tracking is one of the challenging topics in the field of computer vision and target tracking. It has a very wide range of application in many areas, such as military guidance, security monitoring, medical images and meteorological analysis, and the study of the tracking algorithm has important practical significance and theoretical value. In this paper, the complex environment of low-flying targets tracking algorithm is for modern warfare in the low-altitude penetration. Low-altitude penetration is nothing more than evading radar tracking or other detector, to catch the opponent off guard by sudden attack, and then hit targets. The method proposed in this paper is the use of distance measurement sensors, angle measurement sensors and video sensors to obtain target motion information, and to put forward an image fusion target tracking algorithm, which makes the city’s low-flying targets tracking accurately, reliably and steadily. The content of this paper are as follows:Firstly, observation platforms from a single station still start in the fusion three-dimensional space target tracking and two-dimensional image tracking algorithms the basic tracking principle based on the proposed target tracking algorithm. The so-called image fusion, that is, he denotes that in one way, in linear filtering prediction algorithm, it feeds the predicted value of subject under three-dimension coordinate system back to the window coordinate system by a inter-conversion formula, providing a basic for feature matching algorithm; while in another way, two deflection angles which is obtained after successful subject matching, are transformed to tracking axis deflection angles, guaranteeing that measuring devices can get an accurate position information about the subject. This paper also introduces us how to using image preprocessing algorithm to weaken complex background, use multi-frame difference multiplication mean to segment moving target, use clustering algorithm mean to split out the low-flying targets, and extract the characteristics of the moving target.Secondly, based on the single station stationary observation platform, we extend our research to multi-station tracking system. The algorithm proposed in this dissertation solves the problem of the long-time occlusion caused by objects in dual station tracking system. We also introduced the time unify, sensor information transfer and the station layout principle in the multi-station tracking system.Since the experimental environment simulation is difficult to achieve, this paper puts forward a way which uses the toolbox of SIMULINK3D Animation to build a virtual tracking environment. Using the tracking environment to generate video file, and call the Open Source Computer Vision Library and Visual Studio2008platform to design algorithm, to complete tracking algorithm. The simulation results show that the algorithm can make up for the lack of image processing algorithms in object occlusion, and can increase the accuracy and speed of target tracking.Finally, this paper makes a detailed summary of the work and further outlook of the research project.