期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2015
卷号：6
期号：6
页码：4882-4887
出版社：TechScience Publications
摘要：In the past, the plants have been studied for various reasons whether it is for medicine or diseases classification. A plant can be classified on the basis of its flowers, size and shape of the leaves including color, barks etc. The role of leaves in taxonomic classification of plants is very essential. This paper studies the effect of leaf image quality improvement by sharpening and enhancing contrast of the image on the classification accuracy of texture based features. For classification purpose, K-Nearest Neighbor (KNN), J48, Classification and Regression Tree (CART), and Random Forest(RF) algorithms have been used. This study has observed pixels from one pixel distance to five pixel distance in the Co-occurrence matrix for finding the improvement in classification accuracy.
