Carbide image segmentation algorithm based on morphology

Wenyu XIAO,Qizhi TENG,Haibo HE
2015-01-01
Journal of Computer Applications
Abstract:Since the traditional image segmentation method could not excellently resolve the object conglutination problem on carbide segmentation, a method based on morphology, image marker, threshold segmentation and edge extraction was proposed. A modified edge extraction algorithm to eliminate the image noise produced in the procedure of edge enhance was introduced. Then the method of edge extraction and repair based on mathematical morphology and image marker was expounded. In the experiment, firstly, threshold segmentation was used in the target extraction. Then, weak edge points were extracted by using the modified edge extraction algorithm. Last, the distance and direction judgment criteria was introduced to complete the edge connection and repair. The result shows that the method has good effect to resolve the object conglutination problem on carbide segmentation.
What problem does this paper attempt to address?