MACHINE LEARNING FOR SUSTAINABLE AGRICULTURE: ENHANCING PADDY LEAF DISEASE DETECTION USING CONVOLUTIONAL NEURAL NETWORK

Abstract

Sustainable agriculture, crucial for long-term food security, faces challenges in maintaining rice yields amidst growing global demands, particularly in regions like Malaysia where rice is a staple. Early and accurate detection of paddy diseases is vital to minimize crop losses, but traditional manual inspection methods are time-consuming and often inaccurate. This paper addresses the need for a more efficient solution by using Convolutional Neural Network (CNN) model based on the InceptionV3 architecture to detect four common paddy leaf diseases: Bacterial Leaf Blight, Brown Spot, Leaf Smut and Hispa. The model was trained using datasets from Kaggle, employing data preprocessing and augmentation techniques to enhance accuracy. The results show high accuracy (95%), in classifying the targeted diseases, demonstrating its potential for real-world deployment in automated disease detection systems. This study presents a viable solution for modernizing paddy disease detection and management, offering a scalable tool for sustainable agriculture practices that can reduce crop losses and bolster food security amidst growing global demands.