Research On Vision-Based Detection Algorithm Of Airport Perimeter Invasion

Airport perimeter is a security barrier isolating the flight control area and the outside,which is responsible for the security of the flight control area. But, the currently anti-invasion systems have lots of shortcomings, and existing electronic patrol work is very tough.Therefore, how to reduce existing problems of current anti-invasion systems, reduce airport cost and work pressure have become the keys to airport security. Therefore, an airport perimeter anti-invasion suggestion based on stereo vision will be designed, during which navigation control method of the mobile patrol system, breakage detection algorithm of grid images and tracking method of mobile object are studied.Firstly, navigation control method based on information fusion is designed, which is based on odometer, vision and other sensors. Trustiness models of odometer and vision are built according to their error source, and they are computed timely, based on which the localization algorithm is proposed, assuring the safe patrol of the mobile patrol system.Secondly, a breakage detection method of grid images based on improved Harris corner detection theory is presented. Gray "similarity" is introduced to improve the traditional Harris corner detection process, greatly improving the corner extraction speed. According to the corner distribution, the damaged nodes can be located, and the damage area will be calculated.Lastly, mobile object tracking algorithms based on MHI is studied. SIFT feature matching method is adopted to compensate the dynamic background changing, and according to the current and compensated images, the contour of the mobile object is obtained, thus MHI(motion history map) can be used to detect and track the mobile objects.Experiments have been done on the platform of patrol robot, testifying the validity and practicality of proposed algorithms. The accomplished work has some practical significance to the airport anti-invasion.