MANGO PLANT DISEASE DETECTION USING MODIFIED MULTI SUPPORT VECTOR MACHINE ALGORITHM

Abstract

Identifying the mango plant diseases can be done by visualizing its leaf. Plant disease is one of the main problems in the field in agriculture which leads to waste of time and money. Mango plant disease identification at the early stage prevents loss of money and time to the farmers.  The idea is to identify the disease and take proper measures to avoid the heavy losses in the mango crop yield. Manual process of identification takes lot of time if the field is very large. So, the sample images can be taken and given to the algorithm to detect the plant diseases.  Plant diseases means to observe the leaf patterns of the plant. Health of the mango trees and early detection of diseases is very crucial for the farmers for good yields. By Manual observation it is difficult to judge the mango leaf disease. In these investigations, the SVM algorithm is employed to detect the disease of mango trees. Initially, the training set mango images will be given, where the neurons will updates the weights according to the training set. Later the test images are given and with more accuracy the mango leaf disease will be identified. This investigation is carried out on real-time images captured at NRI Institute of Technology, Vijayawada, Andhra Pradesh, India, comprises of 670 pictures from various mango trees. Infected and healthy images are included in the Database. The experimental results exhibits that the proposed model has the higher detection accuracy than the state-of-the-art methodologies.