MULTIMODAL CROP DISEASE DETECTION: A SYSTEMATIC REVIEW MULTIMODAL CROP DISEASE DETECTION: A SYSTEMATIC REVIEW

Abstract

Multimodal approaches for crop disease detection have gained significant attention due to their ability to integrate diverse data sources for improved accuracy. This review categorizes recent studies into five areas: multimodal deep learning and vision transformers, hyperspectral and remote sensing, thermal imaging and UAV applications, CNN–Transformer hybrids and ensemble methods, and comprehensive reviews. Results indicate that frameworks combining RGB, hyperspectral, and thermal imaging achieve accuracies up to 97.8%, while hybrid CNN–Transformer architectures reach 99.7% on benchmark datasets. Despite these advances, challenges remain in scalability, computational cost, and real-world deployment, highlighting the need for lightweight, explainable, and field-validated models.

Code and data availability

公開論文であることは確認できましたが、現在の公式API・許可済み取得経路では本文を自動取得できませんでした。

No evidence-backed public reproduction asset is currently recorded.