Applied Research Of Wavelet Transform On The Profilometry Of Optical Three-dimension Measurement

The application of three dimensional (3-D) shape measurements is of great significance in the area of machine vision, on-line product inspection, geologic prospecting, etc. Fourier transform profilometry is widely used among all the methods of 3-D sensing. This thesis is one of the important parts of geologic structure modeling experiment, and is aimed to get the accurate outline of the anamorphic object. The key techniques like extraction of effective information calculation of the wrap phase and filtering were investigated in this paper. The main works are as follows:1. In this thesis, the basic tenets and the method of extraction of effective information is analyzed. And the error distribution produced by different frequency domain filtering windows is compared, when noise exists and does not exist. The comparison provides guidance for the selection of window functions.2. In Fourier transform profilometry, the frequency is dispersed and confused when detecting complex objects. In order to resolve these problems, FTP measurement based on wavelet digital filter is presented in the thesis. The modulated grating image is handled with muti-storey wavelet decomposition and the effective frequency component is acquired; the selection of the threshold and wavelet base in the discrete wavelet transform and analyzed. And the measurement accuracy will be improved when the discrete wavelet transform is applied.3. In Fourier analysis, the time-frequency domain characteristic is separate. For this reason, a phase demodulation method based on continuous wavelet transform is proposed. By extracting the phase corresponding to the wavelet ridge, the modulated phase distribution relative to the fundamental frequency is obtained finally. Experiments show that, by this method, satisfactory results can be get.