PROCEEDINGS ARTICLE | September 18, 2014
Proc. SPIE. 9282, 7th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Optical Test and Measurement Technology and Equipment
KEYWORDS: Fringe analysis, Wavelet transforms, Image processing, Denoising, Fourier transforms, Image analysis, 3D metrology, Projection systems, Image denoising, 3D image processing
Fringe projection profilometry (FPP) has been widely used for 3-D surface shape measurement with the features of high accuracy, non-contact and fast speed. In FPP, the phase distribution is extracted from the captured distorted fringe pattern, and the height information could subsequently be obtained by the phase-height relation. In actual measurement, the captured pattern usually contains noises, which will influence the precision of the reconstructed result. In order to increase the accuracy of measurement, noise reduction procedure to these fringe patterns is required. The existing noise reducing methods (such as Fourier transform, Wavelet transform) have certain effect. However, they will eliminate some high frequencies generated by a surface with sharp change and make the image blurring. In this paper, we use Curvelet transform to enhance the accuracy of measurement in FPP. The Curvelet transform has the ability of multiscale and multidirection analysis in image processing. It has better descriptions of edges and detailed information of images. Simulations and the experimental results show that the Curvelet transform has an excellent performance in image denoising and it has a wonderful effect on accuracy enhancement of complex surface shape measurement in FPP.
