一种基于神经网络的激光水下目标探测方法

提出了一种利用多层面神经网络进行激光水下目标检测的新方法．理论分析和实验表明，新方法不仅适用于不同的激光探测系统参数，而且适用于对不同海域情况下的自动目标弱信号的检测

A New NN Detection Method in Underwater Target with Blue-Green Laser

On the basis of the chaotic behavior of the background noise and the local predictability of the chaotic signal, we propose a multi plane neural network method to model the atomsphere/seawater optical channel and detect the target from the received signals. Due to various practical factors, it is difficult to construct ideal atomsphere/seawater channel model by using traditional method. In our approach, each neural plane is built to adapt to different environment parameters. Multi plane NN is constructed to model the optical background noise and to detect target from received signal by using the correlation approach.