Fringe skeleton extraction using adaptive refining

Abstract A procedure for fringe skeleton extraction using adaptive refining is presented. Information from both the original image and a rough skeleton is combined in refining processing so that the resulting new skeleton conforms to the morphology of a cosine distribution and the input pattern. The rough skeleton is obtained by extracting pixels with local maximum grey values. The procedure has been successfully tested on noisy fringe patterns. Different operators for finding local extremum are applied in the testing to explore the capability of the refining program. Being general, automatic and convenient for fringe analysis, the procedure is also useful for processing other line patterns.