Development of an RF-DETR Based Model for Sugarcane Leaf Disease Detection Using Deep Learning Techniques

DOI

10.2991/978-94-6239-789-7_51How to use a DOI?

Keywords

Sugarcane Disease Detection; RF-DETR; Region-Focused Attention; Deep Learning; Precision Agriculture; Object Detection; Transformer

Abstract

Red rot, rust, leaf scald and yellow leaf syndrome are the four major foliar diseases of sugarcane (Saccharum Officinarum), which is a major cash crop in a critical situation throughout the world. Most of the conventional detection methods rely on visual inspection by agronomists, which is labour-intensive and inconsistent. By combining a convolutional backbone for capturing local spatial features with a transformer-based encoder-decoder for long-range contextual modelling, this study proposes RF-DETR, a novel hybrid architecture that achieves superior performance. The key feature of RF-DETR is its learnable region-focused attention module that dynamically filters out irrelevant background regions and enhances discriminative disease-related patches. The focused attention allows accurate bounding-box localisation of lesions even in challenging field conditions. The model is trained using 4200 annotated images across five different categories (red rot, rust, leaf scald, yellow leaf disease, healthy leaves) from agricultural fields and public repositories. A standard object-detection metrics reports mAP@0.5 of 93.1%, mAP@0.5:0.95 of 81.4%, and a mean IoU of 87.6% evaluated using experiments. The suitability of the RF-DETR in precision agriculture applications makes it more competitive than the baseline architectures such as CNN, ResNet-50, and Vision Transformer.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Madhulika Mittal
AU  - Md. Iqbal
PY  - 2026
DA  - 2026/10/08
TI  - Development of an RF-DETR Based Model for Sugarcane Leaf Disease Detection Using Deep Learning Techniques
BT  - Proceedings of theInternational Conference on Innovation and Practices in Artificial Intelligence (ICIPAI 2026)
PB  - Atlantis Press
SP  - 765
EP  - 775
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-789-7_51
DO  - 10.2991/978-94-6239-789-7_51
ID  - Mittal2026
ER  -