Wavelet image processing: Application to the defect detection and dimensionality
Received: 22 June 2012; revised version accepted: 24 July 2012
In the quality control field by artificial vision, the optical imagery techniques based on the interferometry know a continuous development, due to their precision, non destructive character and their capacity of real time processing. In this work, we present the fringe projection technique that is based on the projection of periodic fringes on the studied surface followed by an analysis of the reflected optical image. The analysis result is a phase distribution that informs us on several physical parameters such as distortions. One of the phase estimation methods that are commonly used is that based on the wavelet transform. This method estimates the phase distribution of interest through a local analysis of the optical image and then permits a real time processing in order to have the 3D profile of the illuminated surface. A simulation work has been performed with Matlab to prove the reliability of the adopted method.
Keywords: Defect detection; Non destructive testing; Fringe projection; Wavelet transform; Computer vision; Image processing.