Study on Image Segmentation Algorithm Based on Fuzzy Mathematical Morphology

摘要

Aimed at the noise and exiguity catchment basin of image being easy to result in over-segmentation phenomenon when adopted traditional watershed algorithm to segment the image, the paper explored a sort of new improved image segmentation algorithm based fuzzy mathematical morphology. The method firstly adopted opening-closing algorithm based fuzzy mathematical morphology to smooth the image. Then it computed gradient operators based mathematical morphology. Lastly, it segmented the gradient image based on fuzzy mathematical morphology to get the result. The simulation experiment result showed that it not only can eliminate the over-segmentation phenomenon resulted from traditional mathematical morphological segmentation algorithm and realize the goal separation from the background fully, but also can save image detail more completely when using the new image segmentation algorithm based fuzzy mathematical morphology to segment image. And it explains that the new improved algorithm has a better usability.