Plant Disease Classification Using Machine Learning

In the event of a sharp rise in global population, agriculture strives to provide food to it. In agriculture, the detection and diagnosis of diseases occurring in the plants continues to be the arduous task. That is why it is endorsed to predict the diseases when the crops are in early stage. This work is done to develop and implement a disease prediction system by using different machine learning algorithms and a convolutional neural network. The objective of the paper is to grab the attention among the organisations to employ innovative technologies to decrease the diseases that are persistent in plants. The different approaches of ML and image processing with the algorithm that will provide explicit results to recognize the leaves that are healthy and classification algorithms and techniques so that we come to a result of categorically conclusive factor that the leaf is infected by any disease or not. Firstly, the dataset of leaves is divided into directories based on some features extracted from the leaves. Now, the logistic regression estimates the probability of the leave being healthy, based on the dataset of independent variables. The same dataset is also provided to Neural Networks (NN), Support Vector Machines (SVM) and Naïve Bayes algorithms. All the models are analysed with the confusion matrix and K-Fold Cross validation techniques. The proposed model gave the accuracy of 94% using Neural Networks.