Crop disease presents significant threats to global food security and agricultural sustainability. Traditional monitoring methods, reliant on visual inspections and laboratory analyses, are labor intensive and unsuitable for large-scale implementation. Hyperspectral remote sensing has emerged as a promising tool for operational crop disease monitoring. Here, we provide a broad review, starting with a hyperspectral-based description of observable symptoms of common crop disease and then examining hyperspectral features, including spectral and textural features, pigment light absorption, solar-induced chlorophyll fluorescence (SIF), temporal information, and auxiliary data. We also analyze the algorithms used for disease detection, including traditional statistical methods, machine learning (ML)-based methods, and physically based methods. The review highlights the effectiveness of these methods in distinguishing various stressors, detecting early disease, assessing crop resistance, and monitoring large-scale disease. Additionally, we present two case studies of uncrewed aerial vehicle (UAV)-based hyperspectral imaging for maize leaf spot monitoring. Based on a quantitative literature review, we summarize current research trends. Future research should emphasize integrating physical models with deep learning (DL), ensuring the sensitivity and robustness of spectral features and promoting international data sharing.
Maize leaf spot is a common disease that hampers the photosynthesis of maize by destroying the pigment structure of maize leaves, thus reducing the yield. Traditional disease monitoring is time-consuming and laborious. Therefore, a fast and effective method for maize leaf spot disease monitoring is needed to facilitate the efficient management of maize yield and safety. In this study, we adopted UAV multispectral and thermal remote sensing techniques to monitor two types of maize leaf spot diseases, i.e., southern leaf blight caused by Bipolaris maydis and Curvularia leaf spot caused by Curvularia lutana. Four state-of-the-art classifiers (back propagation neural network, random forest (RF), support vector machine, and extreme gradient boosting) were compared to establish an optimal classification model to monitor the incidence of these diseases. Recursive feature elimination (RFE) was employed to select features that are most effective in maize leaf spot disease identification in four stages (4, 12, 19, and 30 days after inoculation). The results showed that multispectral indices involving the red, red edge, and near-infrared bands were the most sensitive to maize leaf spot incidence. In addition, the two thermal features tested (i.e., canopy temperature and normalized canopy temperature) were both found to be important to identify maize leaf spot. Using features filtered with the RFE algorithm and the RF classifier, maize infected with leaf spot diseases were successfully distinguished from healthy maize after 19 days of inoculation, with precision >0.9 and recall >0.95. Nevertheless, the accuracy was much lower (precision = 0.4, recall = 0.53) when disease development was in the early stages. We anticipate that the monitoring of maize leaf spot disease at the early stages might benefit from using hyperspectral and oblique observations.
The accurate extraction of cultivated land information is crucial for optimizing regional farmland layouts and enhancing food supply. To address the problem of low accuracy in existing cultivated land products and the poor applicability of cultivated land extraction methods in fragmented, small parcel agricultural landscapes and complex terrain mapping, this study develops an advanced cultivated land extraction model for the western part of Henan Province, China, utilizing Gaofen-2 (GF-2) imagery and an improved U-Net architecture to achieve a 1 m resolution regional mapping in complex terrain. We obtained optimal input data for the U-Net model by fusing spectral features and vegetation index features from remote sensing images. We evaluated and validated the effectiveness of the proposed method from multiple perspectives and conducted a cultivated land change detection and agricultural landscape fragmentation assessment in the study area. The experimental results show that the proposed method achieved an F1 score of 89.55% for the entire study area, with an F1 score ranging from 83.84% to 90.44% in the hilly or transitional zones. Compared to models that solely rely on spectral features, the feature selection-based model demonstrates superior performance in hilly and adjacent mountainous regions, with improvements of 4.5% in Intersection over Union (IoU). Cultivated land mapping results show that 83.84% of the cultivated land parcels are smaller than 0.64 hectares. From 2017 to 2022, the overall cultivated land area decreased by 15.26 km2, with the most significant reduction occurring in the adjacent hilly areas, where the land parcels are small and fragmented. This trend highlights the urgent need for effective land management strategies to address fragmentation and prevent further loss of cultivated land in these areas. We anticipate that the findings can contribute to precision agriculture management and agricultural modernization in complex terrains of the world.
Soybean, a globally important food and oil crop, requires accurate estimation of above-ground biomass (AGB) to optimize management and prevent yield loss. Despite the availability of various remote sensing methods, systematic research on effectively integrating canopy height (CH) and spectral information for improved AGB estimation remains insufficient. This study addresses this gap using drone data. Three CH utilization approaches were tested: (1) simple combination of CH and spectral vegetation indices (VIs), (2) fusion of CH and VI, and (3) integration of CH, VI, and growing-degree days (GDDs). The results indicate that adding CH always enhances AGB estimation which is based only on VIs, with the fusion approach outperforming simple combination. Incorporating GDD further improved AGB estimation for highly accurate CH data, with the best model achieving a root mean square error (RMSE) of 87.52 ± 5.88 g/m2 and a mean relative error (MRE) of 28.59 ± 1.99%. However, for the multispectral data with low CH accuracy, the VIs + GDD fusion (RMSE = 92.94 ± 6.84 g/m2, MRE = 30.08 ± 2.29%) surpassed CH + VIs + GDD (RMSE = 97.99 ± 6.71 g/m2, MRE = 31.41 ± 2.56%). The findings highlight the role of CH accuracy in AGB estimation and validate the value of growth-stage information in robust modeling. Future research should prioritize the refining of CH prediction and the optimization of composite variable construction to promote the application of this approach in agricultural monitoring.