Research On Image Processing Algorithms Based On Mathematical Morphology

In this paper,image processing algorithms are researched based on mathematical mor-phology. PCB board components'detection algorithms is firstly presented based on basicbinary morphological transforms, then three di?erent methods are presented to do thehydrophobicity image of silicone rubber surfaces segmentation. One method is based onbinary morphology, which uses histogram equalization and mean filtering in the prepro-cess, then does the segmentation based on entropy, and at last filters the results usingbinary reconstruction by opening.This method is simple and e?ective. Another methodis based on gray-scale morphology, which does filtering based on gray-scale morpholog-ical reconstruction before segmentation, and gets smooth edge using Fourier descripterafter segmentation. This method proves to be e?ective for water drop images. The thirdmethod is based on marker-controlled watershed transform, which uses the same methodas method one to get inner marker and outer marker, then modifies the gradient imageand does the watershed transform. The segmentation result is exact for water drop im-ages. The results suggest that each algorithm has its advantage in some respect, and allprove to be e?ective.