An Efficient Approach for Tree Digital Image Segmentation

This paper proposes an improved method to segment tree image based on color and texture feature and amends the segmented result by mathematical morphology. The crown and trunk of one tree have been successfully segmented and the experimental result is deemed effective. The authors conclude that building a standard data base for a range of species, featuring color and texture is a necessary condition and constitutes the essential groundwork for tree image segmentation in order to insure its quality.