Research On The Detection And Tracking System Of Missile Target Based On Image Sequence

With the continuous development of information technology, compared with the traditional conventional weapons and ammunition, precision guided munitions with its high accuracy, powerful, mobile, and other advantages, gradually appreciated by the countries of the world. In such a case, the television guidance is a kind of precise guidance technology, which has important practical significance to the research of the target detection and tracking system. The principle of target detection and tracking algorithm based on image sequence is discussed and studied, and the image processing algorithm is verified by embedded system.As a photoelectric device, the collected video can produce a large amount of image noise.In view of the principle of image noise generation, three effective methods of image denoising, mean filter, median filter and Gauss filter, are deeply studied.According to the image quality degradation caused by the illumination change or other environmental factors, the paper introduces the algorithm of image enhancement based on histogram equalization.The principle and implementation method of three kinds of main moving target detection algorithms are analyzed and discussed, which are adjacent frame difference method, optical flow method and background subtraction method.The foreground extraction method of moving object based on the mixed Gauss background modeling is introduced.The Studio Visual and OpenCV are used to simulate the frame difference method and hybrid Gauss background modeling algorithm in the upper computer.By comparing the simulation results, frame difference method is relatively simple and the algorithm runs fast, and provides an effective algorithm for the later stage of embedded transplantation. The hybrid Gauss background modeling algorithm has a better effect on moving target detection. In addition, in order to better extract the complete moving target area, mathematical morphology method is adopted to deal with the two value image after moving object detection.Moving target tracking algorithms based on region, feature, model and profile are introduced in detail.In-depth study of the Meanshift algorithm and the continuous adaptive mean shift algorithm in the target tracking application. At the same time, the two algorithms are used in the upper computer to track the target in the same video image. Experimental and simulation data show that Camshift algorithm can achieve better real-time tracking of moving objects. Algorithm output centroid coordinates of the tracking target in the image to facilitate the conversion of the image coordinates and the actual coordinates.At last, we design a simple target detection and tracking system and transplant the image processing algorithm to the embedded platform and verify the feasibility of the algorithm.