Plant Classifier Based on Leaf Shape and Texture Features

Abstract

The recognition of plants has become an active area as per the research many plants taxonomic group are at hazard of extermination. To preserve the endanger species new technique are been developed by the researchers. With the great advancement in the technology these days become the key factor for the identification the product and preserve that product with the help of mobile. The area of recognition of plants is trending nowadays because of the fact there are many more plant that need to be identify. More and more advancements are in progress to quickly and efficiently identify the plant. Today’s world is all about spe ed and accurac y of the results. In this thesis , an efficient method of learning is used for the p urpose of classification. In this thesis , we are using two approaches to identify the leaf and both approach consists of three Phases such as pre - processing, extraction of classification and execution phase . And of them first approach is shape based and second approach is texture based. The pre - processing phase involves a representation of image processing steps such as grayscale transformation and limits impr ovement. The main contribution of shape based approach is the p robabilistic neural network (PNN) classification for the efficient recognition of the leaves whereas in te xture based approach support vector m achine (SVM) classify efficient recognition of lea ves. In shaped based approach 13 sheet characteristics that they are extracted and orthogonalized into 6 main variables given as an input vector to the different methods and in texture based local binary p attern (LBP) use to extract features depending upon radius and nearest neighbour.