Reverse engineering from uni-directional CMM scan data

The design and manufacture of sculptured surfaces using the reverse engineering (RE) approach is popular in industry. RE requires the definition of surfaces, a process that is usually performed using a coordinate measuring machine (CMM). However, the present methods for surface model reproduction require a high density of data points, resulting in increased manufacturing times. This paper proposes a method to reduce the manufacturing times. In this method constrained least-square fitting techniques are employed. The splines are then blended to a surface model by a spline segment matching and subdividing algorithm. A model refinement method is proposed, that uses only a fraction of the data points. Experimental results and future developments are also included.