Panoramic Moving Target Detection And Tracking Based On Mobile Robot

With the progress of society and science and technology, video monitoring is widely applied in military and civil fields. General camera vision is narrow, it is difficult to complete monitoring external environment at the saame time. Because the panoramic vision can simultaneously monitor inside the 360 - degree view of information, very good make up for the inadequacy of ordinary camera vision narrower. With the accelerated aging of our society,moving target detection technology can be used to monitor the living conditions of the elderly living alone person. At the same time, the target detection and tracking technology in the modern war can also be used to intercept the missile. In many practical situations, the moving objects tend to move faster, and be covered by other objects. Therefore, the fast and effective detection and tracking of moving objects in video surveillance is a difficult problem, this paper made the following work around this problem:Firstly, in this paper, the commonly used target detection method is done in-depth research. Mainly studied the continuous interframe difference method, respectively to study the two frames and three frame difference method. At the same time also to done in-depth research background difference method, based on the above several methods to detect the effect to do a comparative analysis. In order to highlight the target area of testing,of morphological filtering and connectivity analysis is also studied.Secondly, for the target tracking algorithm is presented in this paper. Through theory deduction, expounds the principle of tracking algorithm. Given the tracking when using the traditional algorithm, the window size is fixed in the tracking process. Due to the target in the process of movement, the size of the image is not set in stone. If we just use only the window with fixed size, it is inevitable to introduce a large amount of background information or not complete all the features of the target extraction. It can cause a big interference for the extraction of target characteristics. Therefore, in order to improve the accuracy of tracking we introducing the adaptive window change strategy.Third, given the panoramic camera assembly used in the study, in a panoramic image of circular cylindrical block there are three areas in the field of vision, when goals are cover, use a separate algorithm is difficult to achieve accurate tracking, in this paper, through the introduction of the filter, to use its prediction can better solve the camera Angle appear larger mask, the use of algorithm to track the effect is not ideal.Finally, on the basis of the study, through the tracked mobile robot platform, build experimental environment for detection and tracking. At the same time, test the method in the detection and tracking if effect is ideal...