Plant Disease Detection And Prevention Using Deep Learning

Identifying disease in plants is a major research topic in agriculture and it is also important so that no crops are going to waste. As India has a large population, it becomes more important to fulfill the increasing demand and for enhancing the yield of crops we need to save the crops which are affected by the diseases. We can find the disease of the plants by classifying the type of disease the crop has and then we can apply prevention methods. It is done by some experts by using some chemicals to find the disease and it is also time consuming and costly. If we can make an automatic system which can detect the disease by analyzing plant leaves then we can make this process cheaper as well as faster with respect to time taken for finding the disease in crops. For that we can use deep learning models which are used to classify the leaves based on disease they have. In this paper, we are finding the disease on the cassava plant. Because people also eat the leaves of this plant. The aim of this paper is to increase the accuracy and efficiency in comparison to the previously proposed models. Deep learning models are more effective than the machine learning models which we got to know while reading about the work which was done previously on plant disease detection. So that is why we used a deep learning model. In this, we used EfficientNet-B0 with k-fold cross validation. EfficientNet is known as best classification models because they are faster and scale on width, depth, and resolution. Our model gives accuracy of 96.68 on the dataset which we collect from kaggle.