Cloud-Based System for Supervised Classification of Plant Diseases Using Convolutional Neural Networks

摘要

Agriculture is a major contributor to India's GDP. A lot of farmers reside in remote areas of the country and are in constant conjecture as to whether their crop is healthy or not. This makes them use pesticides, artificial fertilizers as a precautionary measure, which is detrimental to its edible fitness. This paper aims to address this issue by providing a cloud based system to determine if the crop is healthy or unhealthy. It aims to perform real time classification using supervised learning, along with elucidating the architecture of the system which classifies images of diseased plants, using Convolutional Neural Networks. It highlights the necessity of real time data gathering. The other obstacle this paper aims to mitigate is to reduce the misclassification rate by comparing classification algorithms(CNN and SVM) and choosing the appropriate algorithm. The dataset for the project is gathered from farms in India. The work also intends to provide this dataset of labelled plant diseases for future use.