Crop Disease Prediction Using Deep Learning Algorithms

Abstract

Good crops yield good food which in turn nourishes the human body and mind. But these crops across the globe face the threat of various diseases that remain unidentified, leading to poorer quality and quantity of crops. But, in recent times, the increasing adoption of smartphones worldwide and current developments in image processing of computers enabled by deep models of learning have made smartphone-based disease detection possible. In this chapter, the authors train a deep convolutional neural network (CNN) model to recognize 18 crop species and 28 diseases by feeding it a pre-available dataset of 70,296 photographs of unhealthy and healthy crop leaves taken under control. On a sustained test set, the trained model shows up to 99% accuracy, proving the feasibility of the method.