In visual inspection, images are acquired and analyzed to assess product quality. In particular, for large-scale products, the image acquisition process becomes extensive. Therefore, it is necessary to be automated. Additionally, positioning errors can arise in robotic systems used for image acquisition, and the target product may not always be placed in an ideal position. Therefore, compensating
... [Show full abstract] for these errors is essential for achieving fully automated image acquisition. To address this issue, this study develops a system that automatically acquires inspection images with positioning error correction for visual inspection. The proposed system generates an image acquisition path using only an STL-format CAD model as input. Furthermore, it performs positioning error correction by aligning a point cloud generated from the STL data with a point cloud measured using a depth sensor. To verify the effectiveness of the proposed system, a case study was conducted. A depth sensor mounted on the end effector of an industrial articulated robot was moved along the generated image acquisition path. The target object was placed without precise positioning adjustments, and point clouds were acquired from each imaging point. Positioning errors were calculated, and corrected positions were commanded to the robot. The verification results confirmed that the intended visual inspection images were successfully acquired. It demonstrates the effectiveness of the proposed positioning error correction and automated image acquisition system.