Study And Realization Of Pattern Recognition Algorithm Based On Morphological Neural Networks

SAR (Synthetic Aperture Radar) target recognition is crucial to the success of battlefield awareness and has become a very hot research topic. In recent years, radar target recognition has made steady progress in many fields, including feature extraction, target classification and recognition. Some ATR (automatic target recognize) systems have been built and have been successfully used in the areas like ground detecting and precision guidance in spaceborne/airborne SAR.This thesis first reviews the ATR fundamentals and the state-of-the-art development of SAR target recognition techniques. And try to apply morphological neural network for pattern recognition, according to the superiority supplementary of morphology and neural network, a research to recognition algorithm about morphology neural network is presented here and the morphology structuring elements are optimized. Two groups of data from the MSTAR are used to confirm the validity of the algorithm. Finally more research morphological neural network via changing classification network to validate validity of the algorithmic, the experiment attains good result.The morphological elements (weights of the neural network) of feature extraction layer are trained based on priori knowledge and realize morphology structural element optimization. So the feature extraction to the object is more accurate. The feature extraction and pattern recognition in one network is realized and the recognition rate is improved.