Unwrapping phase map by estimating regions containing phase jumps

We present herewith a phase unwrapping technique by estimation of regions containing 2pi phase jumps based on their phase-distribution features. The more the neighbors around a pixel have phase values near pi and -pi and the smaller the number difference between the neighbors near pi and -pi is, the more possibly the pixel locates near a phase jump. The pixels with high possibility forms the phase-jump regions, which can then be eliminated by compensating the pixels with a multiply of 2pi to minimize the phase difference between them and their neighbors. After removing phase jumps, one can utilize ordinary filtering operations to postprocess the noisy unwrapped phase map. Computer simulation and experimental results from the surface profile measurements of a transparent plate are presented.