A Wavelet Edge Detection Technique Based on Method of Minimum Interclass Variance and Fuzzy Algorithm

李牧,臧希喆,闫继宏,赵杰
DOI: https://doi.org/10.3321/j.issn:0372-2112.2008.09.015
2008-01-01
Abstract:Whatever traditional wavelet edge detection technique was adopted,threshold was needed to filter candidate edge points for edge detection.However,threshold was obtained through experience presently,the best result was received after cut and trial had been used repeatly,in addition,single-threshold calculation method can not accomplish accurate edge detection by now.These shortcomings restrict application of wavelet edge detection technique in practice.Aim at this problem,an algorithm of self-adaptive calculating double thresholds based on method of minimum interclass variance was proposed,and any parameter was not needing artificial setting.Fuzzy algorithm was adopted to link edge pixels,Theory and experiments showed that the algorithm was of validate and correct.
What problem does this paper attempt to address?