Research On Three-dimensional Surface Profilometry Using Spatiotemporal Phase Shifting

Accurate 3-D shape measurement is very useful in many fields, such as producing monitoring, designing and manufacturing, reverse engineering, computer vision and so on. Among so many profilometries, the method of spatiotemporal phase shifting has advantages in accuracy, speed, non-contact over others, so a lot of researchers work hard on this technology. This research is about three-dimensional surface profilometry using spatiotemporal phase shifting with MEMS mirr or. It contains the principle of phase shifting profilometry, the method to project fringe pattern with a MEMS mirror, how to calculate 3D coordinate with the pictures, how to get best values of system parameters through numerical simulation and the experiment of 3-D measurement.First of all, this paper described the achievements about phase shifting profilometry both at home and abroad, and compared different way of get fringe pattern, the advantages and disadvantages of them are discussed, showed readers why we should chose MEMS mirror.Secondly, technical principle of phase shifting profilometry was introduced in detail, and specified how to add projection technology applying MEMS mirror into this theory. Based on the theory, we established mathematical model for the detection system, and wrote computer programs that can process photo data we would get using MATLAB.Then, we simulated the every part of 3-D measurement, and compared the Root Mean Square Errors(RMSE) when different system parameters were se t. By this way, we knew how to build our experiment system makes results more accurate.According to the work we did, built 3-D shape measurement system, including MEMS mirror scanning module, image receiving module and synchronous control module. On the basis of measurement theory and equipment performance, we designed the control signal needed. After testing our system is usable, we measured a flat board and a model of human face separately. The measurement of flat board is used to calibrate system parameters and calculate RMSE, the result is about 6%. The result of human face detection showed us the system can distinguish distinctive part of the sample, and the resolution is about a few millimeters. At last, to solve the problems in our experiments, we gave some advice that may reduce errors.