基于分形的BP网络水下目标识别

主要讨论目标回波的分形特征和基于分形的识别方法,并用实际潜艇的回波数据进行了分形特征识别研究。在分析回波信号的时间域波形的基础上,应用随机分形理论,给出基于分形 Brown 运动的回波信号分形特征矢量提取的理论和方法;提取了回波信号的分形特征矢量;进而给出了基于 BP 网络的分类计算方法。计算结果表明,提出的提取水声回波信号目标特征矢量的方法与分类方法切实可行。

Underwater Target Recognition Based on Fractal Feature

This paper introduces fractal feature of underwater target echo and recognition method based on fractal algorithm detailed and researches on fractal feature recognition by means of utilizing factual echo data gathered from submarine.Subsequently the article embarks on analyzing echo time domain waveform,applies random fractal theory to present theory and method of echo fractal dimension based on fractal Brown motion,distills fractal feature vector (FFV)of echo.Furthermore it presents Computing Method based on Error Back Propagation network.The results of simulation show that the method of underwater echo target feature vector and the method of fractal basically satisfies the technical requirements.