Segmentation of coconut crop bunch from tree images

Harvesting is one of the very crucial stages in crop management. Harvesting the crop at proper time will enhance the quality. In this paper we segmented the coconut crop bunch from tree image. Different segmentation methods like, Color based K-Means clustering, Marker controlled watershed, Grow-cut and Maximum Similarity based Region Merging (MSRM) are explored. Experimentation conducted using a dataset of 200 images for demonstration. Out of these methods the MSRM provides good result.