Indirect target recognition method for overhead infrared image sequences

Target recognition in infrared imagery is a challenging problem due to the complexity of real-world, especially in the cases that the target is occluded or hided completely, and that there exist lots of objects which are similar to the target. For these cases, an indirect target recognition method is proposed to recognize the target for the overhead image sequences. The proposed method mainly includes four steps: preparing template image of landmark, extracting the real-time scene image, matching landmark and locating the target. Several key technologies used in the proposed method are presented, such as perspective transformation used to prepare the template image for matching it with the overhead scene image, position prediction which is proposed to improve the real-time performance, and target localization used for locating the target's position. Experimental results are shown to demonstrate the robustness and efficiency of the proposed method under the condition that the target is occluded and that there exist lots of objects which are similar to the target.