The transition to smart yards is facilitating the automation of the entire manufacturing process, yet the automation of painting operations still presents numerous unresolved issues. Among these, the automation of special painting areas such as the underside of blocks remains a challenge due to the complexity of the painting surface, the presence of obstacles, and the limited distance between the painting surface and the painting robot, posing various issues for practical application. This study, aiming to verify the feasibility of painting operation automation through digital twin simulation, proposes an automation technique utilizing AI-based object and line detection technologies for such specialized painting tasks. By synthesizing images from multiple cameras, extensive data on the painting surface is acquired, and objects and baseline recognition are employed to identify painting surfaces and welding lines, among other painting baselines. This approach has enabled the creation of highly accurate painting operation paths. Moreover, a 3D-based digital twin simulator was developed to verify the painting automation technique.