舰船尾流的激光探测是一种新的鱼雷制导手段。水体的后向散射光信号是舰船尾流后向散射光信号检测的常见干扰,由于其在频域十分接近且信号强度大于尾流信号,因此难以用传统方法提取有用的尾流信号。针对这一问题,提出了一种基于盲源分离的处理方法,将核独立成分分析技术应用于舰船尾流后向散射光信号的提取。介绍了核独立成分分析的基本原理和具体算法,进行了仿真计算,并与传统独立成分分析算法进行比较。结果表明在盲源信号分离中,基于核空间的独立成分分析与其他独立成分分析算法相比更具有准确性。最后应用该方法对海上实验数据进行处理,提取出了舰船尾流信号,取得了良好的效果,验证了该算法的有效性。 The wake detection of the ship is a new means of torpedo guidance. The backscattered optical signal of water body is a common disturbance detected by wakes backscattering optical signal. It is difficult to extract the useful wake signal by traditional methods because it is very close to the frequency domain and the signal intensity is greater than the wake signal. Aiming at this problem, a method based on blind source separation is proposed, and the kernel-independent component analysis (ICA) is applied to the extraction of wake-backscattered optical signals. The basic principles and specific algorithms of kernel independent component analysis are introduced. The simulation calculation is carried out and compared with the traditional independent component analysis algorithm. The results show that independent component analysis based on kernel space is more accurate than other independent component analysis algorithms in blind source signal separation. Finally, this method is applied to the experimental data at sea to extract the ship wake signal and achieve good results. The effectiveness of the proposed algorithm is verified.