The Algorithm Research Of Auto-recognition Of Targets Based On TDSDM642

With the deleveloping of image theorys and DSP application technology, the digital image processing technology of combining together with DSP technology is widely used in every department of national economy, and the requirement of the modern war is higher and higher, the airfight is a complex, dynamic environment even more, so it is more and more important to recognize the sky goal. This paper mostly has discussed recognition technique of plane targets at airfight,it mainly includes the following several aspects:Firstly,analyse study actuality and development direction of this task. This paper mainly studies DM642 software and hardware, procedure of structure design and utilize EDMA transmission means to solve a large number of bottleneck problem of data transmission while pattern process uses.Secondly, this paper mainly put forward different grey level threshold value segmentations of plane target after brief analyse the smooth and sharp image of image strengthen technology, segmentation for the overall situation, adopt the maximum variance method, the maximum entropy combine with adaptive threshold selection method, the maximum variance ration between two classes and in two classes; segmentation for the part situation , has adopted the adaptive threshold value method; To background more complicated segmentation, have adopted the two-dimentional maximum entropy method. Do some improvement to these segmentations and in operation optimize at the speed.These segmentations all get good effect to make preparation for the following recognitiong.Thirdly, extract the feature of the image.In this paper,Hu moment feature is extracted,normalized and modified.Then, moment feature which is input in artificial nerusl network is used to recognize aircraft.In this paper, mainly improved back-propagation and Radial Basis Function recognition methods.Also do some improvement to these method and analyse the result.Next, explain the subject procedure of programme optimize, point out some problem to go on to optimize.Finally, summarized the content of a subject briefly, and looked forward to the hardware and direction of study on this subject of software respect.