Segmentation Method for Rock Particles Image Based on Improved Watershed Algorithm

摘要

The complex texture of natural background gives lots of difficulty to segment foreground objects from the nature background in rock particles image. An improved watershed segmentation method is proposed to raise the segmentation accuracy of rock particles image. The new method combines the merits of mathematical morphology algorithm and traditional watershed algorithm. At different stages of segmentation, this method uses morphological reconstruction, marker computations, defined intuitive watershed algorithm and region merging to avoid over-segmentation and guarantee a better ability of segmentation.