Research On Technology Of Ground Target Matching Recognition In Infrared Downward-looking Police Aircraft

Police aircraft, which is widely applied in both military and civilian applications,has become one of the most popular research topics in many institutions andorganizations for a long time all over the world. The image processing system, which ismainly responsible for detection, identification and location of the target, is anextremely important part of the police aircraft system.To solve the problem of ground target recognition in infrared downward-lookingPolice aircraft, line features from the infrared and visible light images are firstlyextracted and then matched in the Hough parameter space. The main work of this paperis as follows:(1) By analyzing the characteristics and difference of IR image and optical images,we find that the edges are most reliable features in both images and thus adopt edges asthe matching features. In order to facilitate the detection of edges and extraction of linesin the following step, a multi-threshold image pretreatment method based on limitedRidgelet transform is adopted to improve the contrast of images and filter the noise.(2) By analyzing the pros and cons of the commonly used line detectors, a novelline detection algorithm is adopted. This algorithm fitted the salient points on the edgeinto straight lines under the constraints of Helmholtz principle, which can extract thesalient lines in the images effectively.(3) The extracted straight lines is matched by finding the cumulative peak in theHough parameter space, which decompose the similarity transformation parameters intorotation, scale and translation and solve them step by step.Experimental results show that the proposed ground target recognition methodbased on registration of line images can meet the requirement of the tasks of armedpolice, such as patrol duty, maintenance of stability, counter-terrorism and so on.