Study on a gradient algorithm-based method for alignment of ship blocks and virtual blocks in augmented reality

초록

Currently, shipyards are manufacturing ships using CAD-based design drawings, and also using drawings for shipbuilding and outfitting inspections. Recently, the number of skilled workers who can accurately interpret drawings is gradually decreasing in the field, and relatively inexperienced new workers or foreign workers are entering the field. This change leads to a decline in drawing interpretation ability, and causes many difficulties in precise tasks such as equipment installation and outfitting inspection. To solve this problem, a worker support system that intuitively understands design information, clearly understands work without drawings, and accurately performs work is required. Therefore, this study aims to utilize augmented reality, a technology that visualizes a 3D CAD-based ship design model on the actual structure that the worker sees at the site in real time to meet this demand. However, in the case of ship blocks, it is difficult to extract feature points in augmented reality because they are complicated due to internal materials, which limits their use. Therefore, in this study, to solve this problem, we conducted a study on how to align a 3D model of a ship with an actual block by utilizing an image processing technique using images received through a camera. In addition, this study detected one of the internal materials and performed alignment using the extracted features. Finally, validation was performed using the information of features extracted from the aligned 3D model and the actual object for verification.