Detection of Guava Fruit Disease through a Unified Deep Learning Approach for Multi-classification
Raghav Jain, Pulkit Singla, Niharika Niharika, Rishabh Sharma, Vinay Kukreja, Ramanjeet Singh
Abstract
Guava is a popular fruit grown widely in India and is also the largest guava exporter globally. With an annual production of 21.8 million tonnes, ensuring the quality of guava fruit by detecting diseases is crucial. In this study, we propose a deep learning (DL)-based model to classify guava fruit images as healthy or diseased and to identify three specific diseases: leaf spot, guava rust, and guava canker. The model is based on a combination of convolutional neural network (CNN) and long short-term memory (LSTM) algorithms, trained on a dataset of 6000 images. During testing, the model achieved a high accuracy of 95.90%, demonstrating its potential as a valuable tool for disease detection in guava fruit. This research provides a promising approach to accurately detect and classify diseases in guava, which can assist in maintaining high-quality fruit production and improve crop yield.