As a extension of windowed Fourier transform and wavelet transform, S-transform has the features of good time-frequency characteristic and multi-resolution ability. Compared with Fourier transform which lacks local analysis ability, S-transform not only has this ability, its coefficients are also closely connected with the Fourier spectra of the image to be analyzed. Therefore, S-transform has been widely used in fringe analysis in recent years. Based on the features of S-transform coefficients, the 2D S-transform "ridge" (2DSTR) method and 2D S-transform filtering (2DSTF) method are studied deeply. The two methods are compared in detail, too. The results of theoretical analysis are verified by computer simulation and experiment, and they will provide reference for the application of S-transform in 3D shape measurement.