Purpose - The purpose of this paper is to provide a comprehensive work flow for the improvement of the Reverse Engineering (RE) process in producing non-uniform rational B-splines (NURBS) models from scanned point cloud data. This should become a reliable guide in the creation of desired 3D-CAD models in order to improve efficiency of downstream operations and further to make decisions regarding quality control. Design/methodology/approach - The paper deals with a detailed investigation of operations in achieving an object's accuracy in the data editing phase and data fitting phase that employs the use of a 3D scanner. A case example involving the ShapeGrabber (R) AI310 laser scanner was used in digitizing the physical object. Operations considered for investigation at the data editing phase include relaxation, decimation of triangles and sharpening of edges. Contour detection, construct patches, target patch count, grid construction and grid resolution are selected as the operations for investigation in the data fitting phase. Evaluation of the generated digitized models was carried out by performing tests which include 3D Comparisons and Geometric Dimensioning and Tolerance (GD & T) testing. Findings - The process of data editing is considered to be extremely time consuming which requires a high degree of skill in order to carry out the data manipulation steps. For the purpose of investigation, an electrical socket cover was considered as the object for digitization. The study found some contributors to enhance the quality of the digital model that can be used in the first piece inspection. The results indicate that although the operations associated with the data fitting phase affect the overall quality of the digitized model; they are however, limited by whatever the quality achieved at the data editing phase. Practical implications - The RE work flow described in this research will assist designers and practitioners in improving both the efficiency and effectiveness of design and manufacturing functions. Originality/value - The data editing and fitting processes are time consuming due to various adjustments necessary in obtaining a NURBS model from the digitized data. Thus, the proposed RE work flow identified the steps to realize the desired CAD models from the point cloud data. Moreover, from this study, practitioners will get a concise overall understanding about which geometrical features need to be adjusted so that the required model can be achieved; instead of the need to develop this procedure by themselves through the process of trial and error.
