A Survey on Sugarcane Plant Disease Detection Using Deep Learning With Fusion Method

Abstract

ABSTRACT - Sugarcane is one of the most important commercial crops worldwide but its productivity is greatly affected by diseases such as red rot, rust, mosaic, smut and yellow leaf disease. Conventional disease detection techniques are based on manual inspection which is a time-consuming, labor-intensive and error prone process. This paper gives a detailed review of the deep learning methods for the automated detection of sugarcane diseases with a special focus on the fusion methods of stem and leaf features. Different deep learning architectures such as CNN, VGG, ResNet, EfficientNet, DenseNet, MobileNet, and YOLO are analyzed and compared in terms of accuracy, efficiency, and deployment capability. The study also explores multimodal approaches, such as hyperspectral imaging, thermal imaging and environmental data integration, to enhance prediction performance. Reported results show that advanced models like EfficientNet-B7 and DenseNet201 achieve accuracies above 99%, while lightweight models like MobileNet allow for real-time mobile deployment. The review highlights significant research gaps such as small datasets, lack of stem-leaf fusion studies, no severity classification, and real-world deployment issues. Future research directions are related to explainable AI, multimodal fusion, lightweight edge computing models, and precision agriculture applications for sustainable sugarcane cultivation. Key Words: Sugarcane disease detection, Deep learning, CNN, Stem-leaf fusion, Computer vision, Precision agriculture.

Plant phenotyping relevance

サトウキビ病害の画像・深層学習による検出手法を中心にレビューしており、植物の病態を観測・推定するフェノタイピング手法レビューに該当する。

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.