Advancing Paddy Crop Productivity using Deep Learning Ensemble Approach

Maximizing productivity and resource utilization are paramount in paddy agriculture, given its pivotal role in global food provision. One of the foremost challenges confronting paddy cultivation is the prevalence of diseases and pests, which threaten crop health and yield. To address this pressing issue, we have developed an advanced automated system leveraging ensemble deep-learning techniques to detect and classify pests and diseases afflicting paddy crops. The presence of pests and diseases poses a substantial threat, potentially resulting in diminished yields and financial setbacks. We have employed a deep hybrid model comprising Convolutional Neural Networks (CNN) and CatBoost classifier. Our model, trained on real-time images capturing pests and diseased leaves, has demonstrated remarkable accuracy rates of 97.52% for disease prediction and 92.55% for pest identification. This innovative approach enables early detection and diagnosis of ailments and pest infestations, facilitating timely intervention and mitigating production losses. Applying deep learning hybrid models to agriculture increases disease and pest prediction, as well as decision-making, resulting in more efficient and sustainable farming techniques. Farmers and agricultural departments can use the proposed technique to protect yields and boost production in paddy crops.