The Algorithm Research Of Auto-recognition Of Targets Based On TMS320C642DSP

In the modern war, precision air striking and the integrated air-to-ground combat have been the most important combat form, so it is getting more and more important to recognize the sky goal. Directed against the aircraft, this paper has discussed the objective research aircraft based on DSP image processing platform. Following the completion of several major aspects:Firstly, study actuality and development direction of this task are surveyed. 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 studies grey level segmentation of plane target .It is put forward different thresholding methods, including the maximum variance method, the maximum entropy based on Evolutionary Programming, the maximum correlation criterion threshold, iterative threshold; for color image, the adaptive threshold value method based RGB color space has been introduced; for the segmentation of image in complicated background, it provides local threshold. The test shows that these methods can distinguish the object and background perfectly to make preparation for the following recognition.Then, the feature extraction and target recognition algorithm are studied.Hu moment feature and Normalized Moment of Inertia are discussed in this paper .Hu moment feature normalized is applied to the target recognition. Then Radial Basis Function recognition method is adopted, using different kinds of center selection algorithm to make network which is applied to the plane recognition. Analyse and compare the results from the RBF made of center selection algorithm.Finally, study the moving objects detection algorithm, and adopt background subtraction to detect the aircraft target on the basis of analysis of advantage and disadvantage of some mostly used algorithm.