Bridge scour morphology identification and reconstruction using 3D sonar point cloud data

3D multibeam sonar is a feasible solution for detecting bridge scour. However, the reliance on technicians for the identification of the morphological characteristics of local scour pits is time-consuming and subjective, and the absence of surface data hinders scour morphology analysis. Hence, an algorithm is proposed for the unsupervised identification and precise reconstruction of bridge scour morphology. This algorithm segments the scour area using local ternary patterns, optimizes parameters through the dung beetle optimizer, extracts local scour pits with k-means, and introduces an adjustable ball-pivoting algorithm for surface reconstruction by adjusting the mesh ensemble connections. Algorithm testing on the simulated scour data yielded a F1-score of 0.9017 for identification and improved point cloud density, whereas performance evaluation on the Wuhu Yangtze River Bridge in China demonstrated accurate identification of scour morphology and adaptive reconstruction. Thus, the proposed algorithm can enhance the automation and efficiency of scour detection using 3D sonar.