Research On Algorithms Of Infrared Ground Target Detection And Tracking

The detection and tracking of infrared targets are widely used in many fields, such as military, surveillance system, traffic control and so on. As the complexity of the ground background and the uncertainty of interference, it makes the detection and tracking of ground infrared target become the focus of scholars at home and abroad. The thesis studied the detection and tracking of complex ground infrared target. The main achievement of this dissertation is as follows:First of all, the characteristics of ground infrared image are deeply analyzed and studied, an algorithm of suppress noise method is researched and implemented. The widely used methods of infrared target detection and tracking algorithms are researched, when the target is isolated, a new target adaptive detection strategy is introduced, which is based on Otsu threshold and duty ratio. For the dark target in the IR image, we present a novel method for detecting bright and dark target in ground infrared target.Secondly, as the ground background is complex and different parts of the target have different infrared radiation, the algorithm based on threshold is unable to get the accurate target information. In order to extract effective target information for target tracking, a novel method based on duty ratio and image texture is presented for adaptive selection of matching template.Thirdly, we select to use the corresponding tracking method according to detected target information in the target tracking phase. If the target is isolated, the tracking method based on adaptive various feature fusion is used to track the target. While the target information is incomplete, we use the image matching method to track the target. We analyze the occlusion problem in target tracking, a method based on fragment template is utilized to track the target.At last, the template updating is researched and analysis. In order to adjust the template with the target appearance changes and occlusion, different template updating strategies are adopted with respect to different states of target in the tracking.