선박 블록 구조의 3차원 점군 데이터 기반 평면 부재 인식 및 용접선 생성 방법

초록

This study presents a method for identifying major planar parts and automatically generating weld lines from 3D point-cloud data of ship-block structures. First, the point cloud is partitioned into a 3D grid, and the main plate and floor are identified through cell-based plane fitting and plane merging. The floor is then separated into left and right parts and reconstructed, and collar plates are extracted using the positional relationships among neighboring points. Components such as longitudinals, stiffeners, and brackets are classified on the basis of distance and directional features relative to the main plate and floor. Weld lines are generated from the intersections of adjacent plane pairs, whereas those associated with collar plates are separately extracted according to their boundary geometry. In addition, weld-line endpoints are refined by identifying the corresponding points and constructing local planes, after which adjacent segments are connected to form continuous weld lines. Experimental results on two ship-block point-cloud datasets demonstrate that the proposed method can robustly identify major parts and effectively generate weld lines that capture the actual geometry even in complex structures.

키워드