Research On Techniques Of Detection And Discrimination Of Ir Object

The technology for the detection and recognition of the objects which is under the infrared low SNR is the key to extend the acceptance distance of guidance system and enhance the capacity of defense system. The thesis is mainly about the detection and recognition of the objects of infrared image sequences including the pretreatment of pictures, the technology upon the detection and recognition of the objects.At the beginning of the thesis, the characteristics of the infrared pictures are discussed, that is, the pretreatment of original pictures based on the need of subsequent treatment. And the thesis puts forward a new method of the treatment of the infrared pictures. Based on the randomicity of infrared picture's noise and the relativity of the movements of object, the potential areas of the pictures are divided by ratio of gray, and then are smoothed in the way of Gaussian weighted filtering.The thesis following the simplest designing thought of the detection algorithm, put forward the detection algorithm of the infrared objects based on similarity of two adjoining frames' optical flow . Just like this, the accordance of the pretreatment of the infrared pictures and the detection of the object in the same complete frame is realized. And the result of the experiments has already proved the stability .After the analysis on the basic feathers of the infrared pictures and the feather of stability and the methods for the target detection, a group of the new characteristic qualities are put forward, and the discrimination of the infrared objects considering the moment invariants of the better bright area.