利用图像声呐检测海洋风电桩基是海上风电场健康运维的重要手段。针对三维图像声呐数据稀疏导致的成像粗糙与离散性显著问题，采用基于空间分布的去噪算法（statistical outlier removal, SOR）、泊松曲面重建等图像处理技术对Echo Scope声呐的三维成像进行优化。根据局部区域定义、统计信息计算、阈值选择检测去除异常点，通过空间网格划分和求解泊松方程提取等值面，对离散数据进行曲面重建，实现了水下桩基的精细化监测。试验结果表明，该技术可提升声呐对水下桩基成像的密度、平滑性，结合高程信息增强显示技术的融合使用，可区分视像歧义，显著改善成像效果。

The detection of offshore wind power pile foundation by image sonar is an important means for the healthy operation and maintenance of offshore wind farms. In response to the significant issues of coarseness and discreteness in imaging caused by the sparsity of 3D imaging sonar data, image processing techniques such as SOR filtering and Poisson surface reconstruction are applied to optimize the 3D imaging of Echo Scope sonar. The method involves detecting and removing outliers based on local area definition, statistical information calculation, and threshold selection. By dividing the spatial grid and solving the Poisson equation to extract isosurfaces, the method reconstructs surfaces from discrete data, achieving refined monitoring of underwater pile foundations. Experimental results show that this technique can enhance the density and smoothness of sonar imaging of underwater pile foundations. The integration of elevation information with enhanced display technology differentiates visual ambiguities and significantly improves imaging results.