Detection and Classification of Disease Affected Region of Plant Leaves using Image Processing Technique

Abstract

Objective: There is the existence of a variety of plants on this earth surface that plays enormous role in human life. But various factors are there that can destroy plant growth like weather conditions, non-availability of accurate resources, plant diseases and lack of expert knowledge to care plants. Statistical Analysis: Plant diseases are one of the major factors responsible for the reduction of plant growth. In the ancient years, it was not easy to detect the plant diseases on time. But in this computing era, digital image processing rapidly developed that it can be used for various real life applications. Findings: In this research work, plant leaf diseases are detected and classified using the image processing techniques. The fundamental steps of image processing and leaf disease detection and final optimization are used in this work. Here, image acquisition is performed by considering RGB colour based disease affected leaf image. Image contrast is enhanced using Histogram Equalization. Image segmentation is performed with K means clustering. Image feature extraction is performed to extract the features of leaf disease symptoms by maintaining Grey Level Occurrence Matrices. Support Vector Machine is used for the leaf disease detection & classification and finally ant colony optimization is applied for the optimization of concept. Applications/Improvements: For the experimentation, dataset of plant leaf affected from bacterial disease ‘Bacterial Blight’ and fungal diseases ‘Alternaria alternata’, ‘Fungal Leaf Spot’ and ‘Fungus Anthracnose’ are considered. The proposed concept is also evaluated by comparative analysis with the existing concepts of SVM and Improved SVM.