Classifying the Severity of Apple Black Rot Disease with Deep Learning: A Dual CNN and LSTM Approach
Rishabh Sharma, Vinay Kukreja, Prince Sood, Abhishek Bhattacharjee
Abstract
Apple diseases cause significant economic losses to the fruit industry every year. Accurate and timely diagnosis of apple diseases is crucial to prevent the disease’s spread and ensure the production of healthy crops. This study presents a novel hybrid model, combining Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks, for multi-class classification of apple diseases. The model was trained and evaluated on a dataset of images of apple leaves exhibiting different severity degrees of black rot disease. The results of the experiments showed that the hybrid model outperformed traditional single-model approaches, achieving an accuracy of 99.02% in the initial severity degree classification of the disease. This demonstrates the potential of combining CNNs and LSTMs to achieve high accuracy in complex image classification tasks, particularly in the field of plant disease diagnosis. The proposed model provides a valuable tool for apple farmers, researchers, and extension workers in the early detection and management of apple diseases.