Fringe-density estimation by continuous wavelet transform.

For many phase extraction algorithms, a priori knowledge of a fringe-pattern density distribution is beneficial for later processing. A fringe-density estimation method based on a continuous wavelet transform (CWT) is proposed. For a one-dimensional signal the instantaneous frequency detected at the CWT ridge is directly adopted as a measure of the local fringe density. For a two-dimensional signal the instantaneous frequency components in both the x and the y directions are detected. Their reliability is evaluated by the CWT coefficient magnitude, based on which an approximate density value is given. The capability for noise reduction and the accuracy of the method are discussed.