Precise Farming Using Convulationalal Neural Networks
Abstract
Farming is the backbone of the Indian economy. Various crop production and cultivation influence the trade and contribute to the total GDP of our country. The harvests that are contaminated by bugs and sicknesses lead to a decrease in agrarian creation. Generally, the farmers observe the presence of the disease by the naked eye. But this process consumes time and it is expensive and inaccurate too. The proposed method using image processing techniques provides better results than the conventional method. This method is proposed with a new approach to classifying the crop’s diseases using a Machine Learning model that is already trained with a set of images. This is carried out with the help of deep learning networks like Convolutional Neural Network through Transfer Learning and the classification process using Support Vector Machine. The growth of the plants depends on the amount of moisture present in the soil. So the moisture is also measured at periodic intervals and sent to Raspberry Pi 3. The level of moisture content is notified through an app that is developed. Apple and Tomato have been picked as the harvest for test testing as this is one of the significant nourishment crops devoured by numerous individuals. Any part of the plant that has to be tested is captured through Pi Camera and gets regularly updated in Google Drive. The testing image is passed to the model and the subsequent classification of the image is obtained. The preventive measures for unhealthy plants are also notified. The novelty of this project is that it can distinguish healthy and unhealthy plants through the model developed using deep learning. This system will ensure effective automated plant illness detection and will speed up the process of classification without any human intervention.
Keywords: Deep learning, Convolutional Neural Network, Disease classification, Support Vector Machine.