Title: Enhancing agricultural sustainability: harnessing deep learning for intelligent paddy grain classification and recommendation

Authors: Swagatika Tripathy; Premansu Sekhara Rath; Dibya Ranjan Das Adhikary; Bijay Kumar Paikaray

Addresses: Department of Computer Science and Engineering, GIET University, Gunupur, Odisha, 765022, India ' Department of Computer Science and Engineering, GIET University, Gunupur, Odisha, 765022, India ' Centre for Internet of Things, Department of Computer Science and Engineering, Siksha 'O' Anusandhan (Deemed to be University), Bhubaneswar, Odisha, 751030, India ' Centre for Data Science, Department of Computer Science and Engineering, Siksha 'O' Anusandhan (Deemed to be) University, Bhubaneswar, Odisha, 751030, India

Abstract: Precision agriculture uses image analysis and deep learning to improve crop management. This work describes an integrated strategy for identifying, categorising, and suggesting paddy grain kinds based on deep learning. A collection of 48,000 pictures from 12 paddy kinds was analysed using DenseNet201, which achieved 97.78% accuracy. Extracted attributes employed in a custom CNN predicted compatible environment, yield per hectare, and growth duration with 98.13% accuracy while offering alternatives due to cost and availability problems. Odisha's paddy varieties frequently encounter classification challenges due to variable local nomenclature, which impacts seed selection and research. A cosine similarity-based recommendation algorithm assists farmers in selecting visually comparable alternatives that are marketable, processable, and have consistent gastronomic quality. It provides replacements during supply shortages, promotes higher-yielding and disease-resistant types, and helps to keep supply chains stable. Combining study and actual farming improves sustainability, economic stability, and indigenous biodiversity protection.

Keywords: precision agriculture and image analysis; deep learning for crop management; paddy grain classification using deep learning; crop recommendation; DenseNet201; CNN; convolutional neural network; cosine similarity; yield prediction and environmental compatibility; sustainable farming.

DOI: 10.1504/IJCBDD.2025.151225

International Journal of Computational Biology and Drug Design, 2025 Vol.16 No.4, pp.392 - 417

Received: 04 Nov 2024
Accepted: 23 Jun 2025
Published online: 19 Jan 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article