The Study And Application Of Infrared Moving Target Detection Base On Video

With the prominent problem of social security, the video monitoring system has almost become a necessary equipment of each crowded regions. Owing to the high-dependency on manipulation, the trad itional video monitoring system has been gradually unable to adapt to the practical needs of smart times, so the presence of intelligent surveillance system is imperative. The realization of intelligent surveillance system is doomed to be complicated and d ifficult, but the high development of computer vision and image processing technology makes this possible.The heart of intelligent surveillance system is target detection and activity recognition. Depending on the imaging modalities, the intelligent surve illance system can be divided into visible imaging monitoring system and infrared imaging monitoring system. This thesis takes infrared imaging technology as its focus, making the computer to sensing and recording the geometrical information of the object in a monitored field, such as outline, coordinate, posture, motion state and so on, and then analyzing object's character and abnormal behavior.The main research work is as follows:1. Researched on principle of infrared imaging, and picked the best image pre-processing method based on the features of infrared image and then improved the noise suppression and image enhancement method.2. Selective analyzed the object detection algorithm based on OTSU and then put forward two improved object detection algorithm.3. Tracked and filtered for the object and forecasted the motion path based on the target feature model. Picked the Kalman filtering algorithm as the filtering method and then inspected its precision. After that, the traditional gray GM(1, 1) was optimized according to integral background value construction method.4. Elaborated the algorism principle of object abnormal activity recognition and then fitted the moving trail based on the acquired data. After that, completed the abnormal activity recognition set in server rooms, warehouse, road, campus with special demands of the safety. The abnormal activity mainly includes incursion and hesitation behavior.In the end, the shortcoming in this work has been summarized for the prospective in the future.