Fourier transform profilometry based on mean envelope extraction

Based on an image pre-processing algorithm, a three-dimensional (3D) object measurement method is proposed by combining time domain and frequency domain analysis. Firstly, extreme points of sinusoidal fringes under the disturbance of noise are accurately extracted. Secondly, mean envelope of the fringe is obtained through appropriate interpolation method and then removed. Thirdly, phase information is extracted by using specific filtering in Fourier spectrum of the pre-processed fringe pattern. Finally, simulated and experimental results show a good property of the proposed method in accuracy and measurement range. The proposed method can achieve 3D shape of objects having large slopes and/or discontinuous surfaces from one-shot acquisition by using color fringe projection technique and will have wide applications in the fields of fast measurement.