Extracting vein of leaf image based on K-means clustering

Leaf is the primary part of a plant and the major site of food production for the plant. Leaf vein extraction and analysis are useful for investigation of leaf and plant structures. In this paper, a vein extraction algorithm based on the K-means clustering is proposed. Using intensity information, K-means clustering is carried out. According to the clustering results, the boundary of the leaf is extracted and leaf images are divided into two types, the uniform illumination leaf image and the nonuniform illumination leaf image. For a uniform illumination leaf image, vein is directly extracted based on the clustering results. However, for the nonuniform illumination leaf image, some mesophylls are removed first, and K-means clustering is then used to extract the vein. The results show that the proposed algorithm can greatly reduce the misclassification error rate.