基于模糊神经网络的地下小目标识别方法研究

The Recognition Method of Small Underground Objective Based on Fuzzy Neural Network

分析了地下小目标特性,通过提出反映识别目标外形特征及埋设深度的隶属度函数,将常规特征信息模糊化,并结合在模式识别方面更快速简便的PNN神经网络,提出了模糊概率神经网络(FPNN)。最后通过实验数据与BP网络相比较,验证了该识别方法的有效性。

The characteristic of the small underground objective is analyzed, and the membership functions about the shape and depth of objective is presented. With these membership functions, the general feature information is transformed into fuzzy information. Then the fuzzy probabilistic neural network （FPNN） is presented. Finally, compared with BP neural network, the validity of the method is testified by experiment.