蓝绿激光雷达进行水下目标探测成像时,由于受到后向散射噪声影响,图像信噪比随着距离增大而急剧下降。为解决此问题,通过分析水下目标图像噪声污染,提出一种图像预处理算法,通过在频域进行噪声抑制及改进型巴特沃斯高通滤波等处理,能够实现目标复原,提高图像信噪比。以自研的条纹管激光雷达完成室内水池水雷探测成像实验,分别对5m和15m水下水雷目标激光强度像进行频域预处理。结果表明,提出的算法能有效滤除后向散射噪声,提高图像对比度,获得较为清晰的目标轮廓像。 When the blue-green laser radar detects the underwater target, the signal-to-noise ratio of the image drops sharply with the increase of the distance due to the influence of backscattering noise. In order to solve this problem, an image preprocessing algorithm is proposed by analyzing the underwater target image noise pollution. By performing noise suppression and improved Butterworth high-pass filtering in the frequency domain, target restoration can be realized and the image signal-to-noise ratio . The self-developed striped laser radar was used to complete indoor pool mine detection imaging experiments, and the laser intensity images of 5m and 15m underwater mine were preprocessed in frequency domain respectively. The results show that the proposed algorithm can effectively filter backscattering noise, improve image contrast and obtain a more clear target contour image.