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.