自组织神经网络在空中目标识别中的应用

Application of Self-organization Network to Air-target Recognition

本文研究了将自组织神经网络用于空中目标识别的方法，并基于此提出一个识别飞机及其诱饵红外图象的方案，提出目标表面的分形特征和熵特征，针对各特征量纲不统一，信息表达方式差异等特点，利用自组织神经网络来实现特征间的有效融合。实验表明，该方法具有较高的稳定性和准确性。

The application of self-organization network toair-target recognition is studied. On this basis, a new method to recognize the infrared images of airplane and its decoy is presented.The fractal and gray entropy are extracted from infrared images,then the HCMNN is used to fuse the information in decision level.Theexperiments indicate that the new method is stable and reliable.