3차원 점군 데이터에 대한 선박 소부재의 경계 및 홀 인식 방법

Abstract

Ships are constructed by welding steel plates into blocks and assembling these blocks to form the hull. The sub-assembly process involves assembling stiffeners on thin steel plates. To auto- mate this process, it is necessary to automatically recognize the workpiece for welding and per- form dimensional quality inspection. This study proposes a method for extracting holes and boundaries of sub-assemblies in shipbuilding for part recognition and dimensional quality inspection using point cloud data obtained from depth cameras or 3D laser scanners. A grid structure is created in the point cloud data to remove noise and extract external boundaries and holes. Noise is identified by relatively low point density areas, and boundaries and holes can be created using points from the grid encountered first when approached from outside or inside. This proposed method has been tested on several real models to demonstrate the feasibility.