Research on Hull Segmentation and 3D Reconstruction Algorithm Based on YOLOv8

With the increasing demand for hull structure inspection and 3D reconstruction in the shipping industry, the traditional manual and rule-based segmentation methods are difficult to cope with the challenges of complex port environments. In this paper, a ship hull segmentation and 3D reconstruction method combining YOLOv8 and an improved RANSAC algorithm is proposed. Firstly, the YOLOv8 algorithm is used to accurately identify the ship's cabin opening in the scene point cloud in the complex port environment. Based on the location information of the ship's cabin obtained from the recognition result, the algorithm can achieve accurate segmentation of the ship's cabin through the improved RANSAC algorithm, and greatly improve the segmentation efficiency. The segmented ship hatches are further refined by the super voxel clustering algorithm to eliminate or reduce the interference of stray points on the 3D reconstruction. Then the cabin point cloud is reconstructed in 3D by scattering outer contour extraction to maximally recover the topology of the cabin. The experimental results show that the method can accurately and efficiently complete the segmentation and 3D reconstruction of the cabin in the complex port environment with large noise interference. Field validation proves the practicality of this paper's algorithm in real-world scenarios of automated loading in ports, which significantly improves the efficiency of port operations. Overall, the solution proposed in this paper provides a new technical path for the automatic identification and reconstruction of ship hull structures in complex environments and has good application prospects.