New Approach for 3D Shape Measurement Based on Color-Coded Fringe and Neural Network

QIN Da-hui,SHI Yu-sheng,WANG Cong-jun,LI Zhong-wei (State Key Laboratory of Material Processing and Die ＆ Mould Technology,Huazhong University of Science and Technology,Wuhan,China)

DOI: https://doi.org/10.19583/j.1003-4951.2008.02.008

2008-01-01

Abstract:A new 3D surface contouring and ranging system based on digital fringe projection and phase shifting technique is presented. Using the phase-shift technique, points cloud with high spatial resolution and limited accuracy can be generated. Stereo-pair images obtained from two cameras can be used to compute 3D world coordinates of a point using traditional active triangulation approach, yet the camera calibration is crucial. Neural network is a well-known approach to approximate a nonlinear system without an explicit physical model, in this work it is used to train the stereo vision application system to calculating 3D world coordinates such that the camera calibration can be bypassed. The training set for neural network consists of a variety of stereo-pair images and the corresponding 3D world coordinates. The picture elements correspondence problem is solved by using projected color-coded fringes with different orientations. Color imbalance is completely eliminated by the new color-coded method. Once the high accuracy correspondence of 2D images with 3D points is acquired, high precision 3D points cloud can be recognized by the well trained net. The obvious advantage of this approach is that high spatial resolution can be obtained by the phase-shifting technique and high accuracy 3D object point coordinates are achieved by the well trained net which is independent of the camera model works for any type of camera. Some experiments verified the performance of the method.