Identification and Classification of Plant Leaf Diseases Using Data Augmentation and Deep Convolutional Neural Networks

Summary

In agriculture, the early and accurate identification of plant diseases is an important step in increasing crop yield. Leaf disease detection is one of the most widely used methods for identifying plant diseases. In recent years, automatic and accurate classification of plant leaf diseases has attracted increasing research attention. With the emergence of deep neural networks (DNN), the identification and classification of plant leaf diseases have achieved competitive results. This study proposes a method based on data augmentation techniques and advanced deep neural network models to improve the accuracy of the identification and classification of plant leaf diseases. First, augmentation techniques using image processing have been applied to efficiently train DNN models. Then, the advanced DNN models were fine‐tuned to improve the classification accuracy and execution time. The method was evaluated on two public datasets (apple and potato leaf diseases) with accuracies of 98.08% and 99.07%, respectively.