Damaged region positioning for flexible remanufacturing using structured light

An approach for damaged region inspection and positioning in a flexible remanufacturing system is presented. The system is based on a welding robot and a measuring robot. A structured light vision sensor is mounted at the end of an industrial robot to serve as a measuring robot to obtain the 3-D surface data of damaged workpiece. The damaged region can be determined by the geometric alignment of a standard CAD model to the obtained 3-D surface data. The aforementioned geometric alignment is achieved by a registration algorithm. The registration process presented in this study consists of initial and precise registration. In initial registration stage, principal component analysis (PCA) is firstly applied to get the main direction of point cloud. After that the iterative closest point (ICP) algorithm is employed to achieve a precise registration of point cloud and standard CAD model. Then the damaged region data of workpiece is determined by calculating the errors between point cloud and standard CAD model. Finally, some real experimental results are given to illustrate the effectiveness of the proposed methodology.