An Empirical Overview for Plant Disease Detection and Severity Level Classification Using Hyperspectral Images

Plant disease detection is crucial for ensuring crop health and productivity. Early identification helps minimize crop losses, enhances food security, and reduces pesticide use, leading to targeted treatments, better yield quality, and reduced environmental impact. This survey provides a comprehensive framework for plant disease detection using hyperspectral leaf images, categorizing techniques into Deep Learning (DL), Machine Learning (ML), and Internet of Things (IoT) approaches. It offers a detailed analysis of each method, identifies research gaps, and highlights limitations in existing techniques to guide future research and model improvements. The study reviews 25 research articles on plant disease detection methodologies. Among these, DL emerges as the most effective approach due to its ability to automatically learn and extract complex features from raw hyperspectral data, reducing the need for extensive manual feature engineering. Python is commonly used for its integration with visualization tools, essential for analyzing hyperspectral data. While DL techniques are effective, they struggle with generalization across different plant species and environmental conditions. Future research should focus on enhancing model adaptability and integrating multimodal data to advance interpretability. This analysis aims to guide researchers in refining plant disease detection methods using hyperspectral imaging for improved accuracy and robustness.