Phase Unwrapping
There are many medical, military and industrial digital image processing applications which as part of their fundamental operation depend upon the extraction of a phase signal from their input image. Examples of such applications are magnetic resonance imaging (MRI), synthetic aperture radar (SAR), fringe pattern analysis, tomography and spectroscopy. These systems use modern and mature algorithms to extract the phase signal. However the phase is returned in a form that suffers from 2-pi phase jumps due to the use of the mathematical arctangent function, which produces an inherently wrapped output. This wrapped phase is unusable until the phase discontinuities are removed and this is achieved practically by employing a phase unwrapping algorithm.
Phase unwrapping is one of the most active areas of research in digital image processing and has been extensively researched in the last twenty years with more than 500 journal papers having been published, all of which suggest various solutions to this problem. The proposed algorithms produce a variety of differing results and no perfect "correct" solution is guaranteed, even if such a solution exists. Instead, all modern phase unwrapping techniques produce estimations of the "correct" solution. On many occasions, these complex algorithms cannot resolve all the phase wraps in the image and many 2-pi phase jumps still exist in the unwrapped phase map. Problems with 2D unwrapping are not limited to a small subset of challenging images; we are talking here about the majority of real images encountered in many practical applications.
Generally speaking, when these algorithms are executed using digital computers, the more robust unwrappers usually require long processing times. Also, the required processing power increases as the noise corrupting the wrapped phase map increases. On the other hand, the faster phase unwrappers are usually not robust and cannot be used for real applications.
In GERI, a significant amount of research has been carried out in order to develop robust and fast two and three-dimensional phase unwrappers. These algorithms have been published in international journals and conferences. An introduction to the specific algorithms that have been developed by the Institute is included here on our website and the code is made freely available for download for non-commercial applications, subject to certain terms and conditions. If you have a commercial application and are interested in including GERI unwrapping algorithms within your product, then this may be accommodated under license. Phase unwrapping algorithms may also be customised in order to meet your application requirements. Please contact Dr. Munther Gdeisat M.A.Gdeisat@ljmu.ac.uk if you have a commercial enquiry regarding use of our phase unwrapping algorithms.
