Abstract: Wheat image segmentation is the basic work of applying computer vision technology to dynamic monitoring wheat growing. This paper proposed an algorithm combining K-means clustering and mathematical morphology which full use of the color differences between wheat plant and background in the image. Firstly, clustering image based on colors, and then mathematical morphology opening operation was used to eliminate noise. Experiment results showed that the algorithm was effective in segmenting wheat color images．

Key words: K-means clustering, mathematical morphology, Lab color space