Digital Image Processing

Basavaraj S. Anami ¹, Suvarna S. Nandyal ², A. Govardhan ³, P. S. Hiremath ⁴

Affiliations
1 K. L. E. Institute of Technology, Hubli, Karnataka, IN
2 Department of Computer Science and Engineering, Jawaharlal Nehru Technological University, Hyderabad-500085, Andhra Pradesh, IN
3 JNTUH, AP, IN
4 Department of P. G. Studies and Research in Computer Science, Gulbarga University, Gulbarga-585106, IN

Abstract

A plant is identified by features such as stem, leaves, flowers, fruits. But these features differ from one plant to the other. Medicinal plants are not the exceptions. The medicinal plants are broadly classified into herbs, shrubs and trees based on their heights. The medicinal plants are recognized by the experts in the domain. Automation of their identification and classification is very useful in the real world. We have proposed a method, wherein a plant image is segmented into two parts, namely, stem and leaves using k-means clustering. The areas of both the parts are estimated. The aspect ratio is computed, which is the ratio of the area of the stem to the spread of leaves of a plant. The aspect ratios form knowledge base and further, a knowledge based classifier is designed for the classification. Experimental results have shown the classification accuracy of 72%, 65%, 52% for trees, herbs and shrubs respectively. The work is useful in the preparation of Ayurvedic medicines, home remedies, and botanical study.

Keywords

Aspect Ratio, Height, Herbs, K-Means Clustering, Medicinal Plants, Shrubs and Trees.

Full Text