Abstract: As the key link of Qiantang River seawall revetment engineering, the detection and identification of underwater block stone is very important to ensure the construction efficiency of sheet piles. Based on the research on the core technology of shore-based communication and data acquisition of the unmanned ship measurement system, the NORBIT multi-beam sonar resolution estimation model and the maximum beam opening angle, the minimum data update rate and the maximum speed of the additional rock particle size constraints are derived. The relationship and recommended value of,combined with the three-dimensional characteristics of the Qiantang River underwater multi-beam scanning point cloud data, the region growing algorithm is improved to realize the identification and extraction of the boulder point cloud. The engineering practice results show that the optimized multi-beam sonar parameters of the unmanned ship measurement system can achieve accurate detection of small underwater rocks, and can faithfully reflect the real situation of underwater block stone and surrounding micro-topography. The point cloud of underwater block stone is extracted from the point cloud data of many riverbeds to provide accurate distribution of boulders for sheet pile construction, which has good engineering application significance.

Key words: revetment engineering, unmanned ship measurement system, underwater block stone, multi-beam, parameter optimization, regional growth algorithm