基于LMBP神经网络的声纳图像识别

Sonar image recognition based on LMBP neural network

利用具有二阶收敛效应的Levenberg-Marquardt(L-M)算法优化BP的权值修正量,构造了快速收敛的LMBP学习算法,并将其应用在声纳图像识别系统中。通过与标准BP算法和几种常用改进型BP神经网络以及径向基函数网络比较,验证了用LMBP神经网络作为声纳图像识别系统中的分类器,能够提高图像的识别率,加快网络的收敛速度,通过对受不同程度高斯白噪声污染的声纳图象的识别,验证了其性能稳定,具有较好的抗噪声性能。

Levenberg-Marquardt (L-M) algorithm which has second-order convergence effects is utilized to optimize standard BP algorithm and it is applied to the sonar image recognition systems. The experimental results indicate that the LMBP neural network has higher classification rate, more rapidly converge velocity by comparing with other improved BP algorithm and the Radial Basis Function neural network. Through the experiments of recognizing the sonar image polluted with Gaussian noise, it is proved that the LMBP neural network has good stability and better anti-noise performance.