Recent advances in optical sensors and information processing technology have granted a rapid progress in three-dimensional (3D) optical metrology, enabling highly accurate 3D inspection of complex-shaped objects. However, the 3D object surface estimation requires a phase recovery from so-called intensity fringe patterns of the tested objects. In this work, we present a comparative study of three most highlighted phase retrieval methods (phase shifting, Fourier transform and wavelet transform) using single or multiple acquired fringe patterns. Moreover, the phase retrieval methods performance is tested on additive and multiplicative noise based degraded fringe patterns, giving some qualitative and quantitative conclusions. Some experimental data are processed using Fourier transform and wavelet transform based algorithms to confirm accuracy of these two phase retrieval techniques.