Sooty mould mango disease identification using deep learning

Abstract

In India, half of the population depends on agriculture to lead their life. Our country is the largest producer of Mangoes. The scientific name of the plant is Mangiferae. Mango plants are affected by the fungus and pests which reduces the quality and quantity of the product. Already farmers are suffering from lot many problems and we have to support them to improve their economy. Our project aims to increase the mango fruit productivity by controlling the plant disease by early identification through deep learning. We have taken a major disease that affects the mango plant in Tamil Nadu-Sooty Mould In places like Dharmapuri and Triuvallur as the varieties of Mangoes such as Neelum, Alphonso, Bangalora are mostly affected by this disease and the yield drops out. Plants infected by Sooty Mould have a velvety coating over the leaves. It is due to the honey dew secretions. The insects stick to leaf surface and lead to fungal growth. But no direct damage is done by the fungus. The photosynthetic activity is affected adversely due to the blockage of stomata. We propose a solution for detection and classification of plant leaf disease in early stage itself. Deep learning constitutes a modern technique for image processing and data analysis. Deep learning technique has lot of applications in agricultural domain. The Deep Learning methodology, CNN model is developed to perform plant disease detection from leaves images.