Robot Vision System and Artificial Neural Network for Model Reconstruction in Reverse Engineering

A novel vision system and model reconstruction scheme is addressed in this article. The main contribution contains two aspects: first, robot with laser scanning head is used for establishing a vision system, this system can measure 3D surface data for complex freeform surface on line at manufacturing site. Then, using artificial neural network (ANN), a model reconstruction algorithm is developed. For demonstrating effectiveness of the presented ANN digital surface model, a group point cloud data with good accuracy, which was measured by the presented robotics vision system for an existing part, is taken as data sample to train the ANN so that the surface model of the object is reconstructed. By comparing the ANN data with the sample value at the same position point, it is shown that the reconstructed ANN model can match the surface data very well