New data processing method for improving the measurement accuracy in FTP

During the 3D shape measurement, there are noises in the images that are obtained by the capture system. The traditional method, Fourier transform profilometry (FTP) technique, improves the accuracy only by the filtering method in the frequency domain. In this paper, the curve fitting method is used for the light field distribution calculation before the filtering process applied in the frequency domain by choosing a suitable filter window, and then the higher quality of the basic frequency component signal is got. This method can avoid the frequency overlapping caused by the noise, so the improvement of the measuring accuracy of FTP is realized.