Wheel Tread 3D Shape Recovery And Defect Detection Based On Wavelet Transform Profilometry

With the rapid development of modern industrial manufacturing technology, three-dimensional measurement have a wide range of demands for industrial automation monitoring, quality control, machine vision, reverse engineering, biomedical engineering, costume design and other fields. Due to advantages such as non-contact, high precision, high speed, optical 3D measurement technology has become one of the research hot spots. Wavelet analysis is a new mathematic theory and method, and so far it has successfully been applied in many fields. Wavelet analysis method is characterized in multi-scale analysis. It can analysis signal or image with high resolution, by any frequency band, at any part of the object, and with any scale of window. Therefore, it is known as "mathematic microscope". In recent years, wavelet transform technology has been applied in 3-D reconstruction to enhance measuring performance. It highly improved the development of dynamic 3-D measurement base on optical method.The theory and application of wavelet transform profilometry were studied in this thesis, the main work of the paper is as follows:First, this thesis studies the basic theory of wavelet transform profilometry. The key technologies of WTP such as the selection of mother wavelet function, ridge extraction, phase unwrapping were simulation experimented by computer. An extraction method using improved maximum value of the ridge, by setting the threshold to define the extraction point to reduce noise interference to the measurement results, and improves the accuracy of extracting ridge, thereby improving measurement precision of the WTP.Secondly, In optical three-dimensional shape measurement technique, either through or through the airspace of phase modulation-phase modulation in the time domain, whether using phase-shifting method is using the method of wavelet transform to transform profilometry, We tend to be obtained through inverse tangent function of phase difference, in order to get a continuous surface morphology, must be evaluated in a phase recovery from the package to get the actual phase distribution. This thesis studies improve on the algorithm by changing branch setting and adding pretreatment. Last we will get the better result by the new algorithm using computer emulation mode.Finally, the thesis use the wavelet transform profilometry to achieve wheels tread defects detection, by the method the surface defects were accurately detected.