Hybrid CNN-RNN Architecture for Accurate Tomato Disease Diagnosis with Xception-Gru

Abstract

Plant diseases and pernicious insects are a considerable threat in the agriculture sector. Therefore, early detection and diagnosis of these diseases are essential. The ongoing development of profound deep learning methods has greatly helped in the detection of plant diseases, granting a vigorous tool with exceptionally precise outcomes but the accuracy of deep learning models depends on the volume and the quality of labeled data for training. In this paper, we have proposed a deep learning-based method for tomato disease detection that utilizes the ensemble convolutional neural network (CNN) and recurrent neural network (RNN) architecture. The proposed model is known as Xception-GRU as it begins with the Xception pre-trained model and followed by the GRU layers. Thereafter, the Xception-GRU model is trained on synthetic and real images using transfer learning to classify the tomato leaves images into ten categories of diseases. Three different classifiers are used on the features extracted from the Xception-GRU model. These classifiers are the k-nearest neighbor (KNN), multi-layer perceptron (MLP), and support vector machine (SVM). The roposed model has been trained and tested extensively on publicly available PlantVillage dataset. The proposed method achieved an accuracy of 100%, 98.79%, 99.85%, and 100% for tomato leaf image diseases classification into (Early, Late Blight, and Healthy), (Late and Early Blight), (Late Blight and Healthy), and (Early and Healthy). The proposed approach shows its superiority over the existing methodologies.