Exploring UAV narrow-band hyperspectral indices and crop functional traits derived from radiative transfer models to detect wheat powdery mildew
The wheat powdery mildew (WPM) is one of the most severe crop diseases worldwide, affecting wheat growth and causing yield losses. The WPM was a bottom-up disease that caused the loss of cell integrity, leaf wilting, and canopy structure damage with these symptoms altering the crop’s functional trai...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-07-01
|
| Series: | International Journal of Applied Earth Observations and Geoinformation |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1569843225002742 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850209083814051840 |
|---|---|
| author | Yang Liu Xiaoyang Ma Lulu An Hong Sun Fangkui Zhao Xiaojing Yan Yuntao Ma Minzan Li |
| author_facet | Yang Liu Xiaoyang Ma Lulu An Hong Sun Fangkui Zhao Xiaojing Yan Yuntao Ma Minzan Li |
| author_sort | Yang Liu |
| collection | DOAJ |
| description | The wheat powdery mildew (WPM) is one of the most severe crop diseases worldwide, affecting wheat growth and causing yield losses. The WPM was a bottom-up disease that caused the loss of cell integrity, leaf wilting, and canopy structure damage with these symptoms altering the crop’s functional traits (CFT) and canopy spectra. The unmanned aerial vehicle (UAV)-based hyperspectral analysis became a mainstream method for WPM detection. However, the CFT changes experienced by infected wheats, the relationship between CFT and canopy spectra, and their role in WPM detection remained unclear, which might blur the understanding for the WPM infection. Therefore, this study aimed to propose a new method that considered the role of CFT for detecting WPM and estimating disease severity. The UAV hyperspectral data used in this study were collected from the Plant Protection Institute’s research demonstration base, Xinxiang city, China, covering a broad range of WPM severity (0–85 %) from 2022 to 2024. The potential of eight CFT [leaf structure parameter (N), leaf area index (LAI), chlorophyll a + b content (Cab), carotenoids (Car), Car/Cab, anthocyanins (Ant), canopy chlorophyll content (CCC) and average leaf angle (Deg)] obtained from a hybrid method combining a radiative transfer model and random forest (RF) and fifty-five narrow-band hyperspectral indices (NHI) was explored in WPM detection. Results indicated that N, Cab, Ant, Car, LAI, and CCC showed a decreasing trend with increasing disease severity, while Deg and Car/Cab exhibited the opposite pattern. There were marked differences between healthy samples and the two higher infection levels (moderate and severe infection) for Cab, Car, LAI, Deg, CCC, and Car/Cab. N and Ant only showed significant differences between the healthy samples and the highest infection level (severe infection). As Cab, Car, and Ant decreased, the spectral reflectance in the visible light region increased. The decrease in N and LAI was accompanied by a reduction in reflectance across the entire spectral range and the near-infrared area, which was exactly the opposite of Deg. The introduction of CFT greatly improved the accuracy of the WPM severity estimation model with R2 of 0.92. Features related to photosynthesis, pigment content, and canopy structure played a decisive role in estimating WPM severity. Also, results found that the feature importance showed a remarkable interchange as increasing disease levels. Using features that described canopy structure changes, such as optimized soil adjusted vegetation index, LAI, visible atmospherically resistant indices, and CCC, the mild infection stage of this disease was most easily distinguished from healthy samples. In contrast, most severe impacts of WPM were best characterized by features related to photosynthesis (e.g., photochemical reflectance index 515) and pigment content (e.g., normalized phaeophytinization index). This study help deepen the understanding of symptoms and spectral responses caused by WPM infection. |
| format | Article |
| id | doaj-art-265bf84cb0b64c298a69c9ba3d5bc6b7 |
| institution | OA Journals |
| issn | 1569-8432 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | Elsevier |
| record_format | Article |
| series | International Journal of Applied Earth Observations and Geoinformation |
| spelling | doaj-art-265bf84cb0b64c298a69c9ba3d5bc6b72025-08-20T02:10:06ZengElsevierInternational Journal of Applied Earth Observations and Geoinformation1569-84322025-07-0114110462710.1016/j.jag.2025.104627Exploring UAV narrow-band hyperspectral indices and crop functional traits derived from radiative transfer models to detect wheat powdery mildewYang Liu0Xiaoyang Ma1Lulu An2Hong Sun3Fangkui Zhao4Xiaojing Yan5Yuntao Ma6Minzan Li7Key Lab of Smart Agriculture System, Ministry of Education, China Agricultural University, Beijing 100083, ChinaKey Lab of Smart Agriculture System, Ministry of Education, China Agricultural University, Beijing 100083, ChinaKey Lab of Smart Agriculture System, Ministry of Education, China Agricultural University, Beijing 100083, ChinaKey Lab of Smart Agriculture System, Ministry of Education, China Agricultural University, Beijing 100083, China; Key Laboratory of Agricultural Information Acquisition Technology, Ministry of Agriculture and Rural Affairs, China Agricultural University, Beijing 100083, China; Corresponding author.State Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing 100193, ChinaState Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing 100193, ChinaCollege of Land Science and Technology, China Agricultural University, Beijing 100193, ChinaKey Lab of Smart Agriculture System, Ministry of Education, China Agricultural University, Beijing 100083, China; Key Laboratory of Agricultural Information Acquisition Technology, Ministry of Agriculture and Rural Affairs, China Agricultural University, Beijing 100083, ChinaThe wheat powdery mildew (WPM) is one of the most severe crop diseases worldwide, affecting wheat growth and causing yield losses. The WPM was a bottom-up disease that caused the loss of cell integrity, leaf wilting, and canopy structure damage with these symptoms altering the crop’s functional traits (CFT) and canopy spectra. The unmanned aerial vehicle (UAV)-based hyperspectral analysis became a mainstream method for WPM detection. However, the CFT changes experienced by infected wheats, the relationship between CFT and canopy spectra, and their role in WPM detection remained unclear, which might blur the understanding for the WPM infection. Therefore, this study aimed to propose a new method that considered the role of CFT for detecting WPM and estimating disease severity. The UAV hyperspectral data used in this study were collected from the Plant Protection Institute’s research demonstration base, Xinxiang city, China, covering a broad range of WPM severity (0–85 %) from 2022 to 2024. The potential of eight CFT [leaf structure parameter (N), leaf area index (LAI), chlorophyll a + b content (Cab), carotenoids (Car), Car/Cab, anthocyanins (Ant), canopy chlorophyll content (CCC) and average leaf angle (Deg)] obtained from a hybrid method combining a radiative transfer model and random forest (RF) and fifty-five narrow-band hyperspectral indices (NHI) was explored in WPM detection. Results indicated that N, Cab, Ant, Car, LAI, and CCC showed a decreasing trend with increasing disease severity, while Deg and Car/Cab exhibited the opposite pattern. There were marked differences between healthy samples and the two higher infection levels (moderate and severe infection) for Cab, Car, LAI, Deg, CCC, and Car/Cab. N and Ant only showed significant differences between the healthy samples and the highest infection level (severe infection). As Cab, Car, and Ant decreased, the spectral reflectance in the visible light region increased. The decrease in N and LAI was accompanied by a reduction in reflectance across the entire spectral range and the near-infrared area, which was exactly the opposite of Deg. The introduction of CFT greatly improved the accuracy of the WPM severity estimation model with R2 of 0.92. Features related to photosynthesis, pigment content, and canopy structure played a decisive role in estimating WPM severity. Also, results found that the feature importance showed a remarkable interchange as increasing disease levels. Using features that described canopy structure changes, such as optimized soil adjusted vegetation index, LAI, visible atmospherically resistant indices, and CCC, the mild infection stage of this disease was most easily distinguished from healthy samples. In contrast, most severe impacts of WPM were best characterized by features related to photosynthesis (e.g., photochemical reflectance index 515) and pigment content (e.g., normalized phaeophytinization index). This study help deepen the understanding of symptoms and spectral responses caused by WPM infection.http://www.sciencedirect.com/science/article/pii/S1569843225002742Wheat powdery mildewUnmanned aerial vehicleRadiative transfer modelCrop functional traitsNarrow-band hyperspectral indices |
| spellingShingle | Yang Liu Xiaoyang Ma Lulu An Hong Sun Fangkui Zhao Xiaojing Yan Yuntao Ma Minzan Li Exploring UAV narrow-band hyperspectral indices and crop functional traits derived from radiative transfer models to detect wheat powdery mildew International Journal of Applied Earth Observations and Geoinformation Wheat powdery mildew Unmanned aerial vehicle Radiative transfer model Crop functional traits Narrow-band hyperspectral indices |
| title | Exploring UAV narrow-band hyperspectral indices and crop functional traits derived from radiative transfer models to detect wheat powdery mildew |
| title_full | Exploring UAV narrow-band hyperspectral indices and crop functional traits derived from radiative transfer models to detect wheat powdery mildew |
| title_fullStr | Exploring UAV narrow-band hyperspectral indices and crop functional traits derived from radiative transfer models to detect wheat powdery mildew |
| title_full_unstemmed | Exploring UAV narrow-band hyperspectral indices and crop functional traits derived from radiative transfer models to detect wheat powdery mildew |
| title_short | Exploring UAV narrow-band hyperspectral indices and crop functional traits derived from radiative transfer models to detect wheat powdery mildew |
| title_sort | exploring uav narrow band hyperspectral indices and crop functional traits derived from radiative transfer models to detect wheat powdery mildew |
| topic | Wheat powdery mildew Unmanned aerial vehicle Radiative transfer model Crop functional traits Narrow-band hyperspectral indices |
| url | http://www.sciencedirect.com/science/article/pii/S1569843225002742 |
| work_keys_str_mv | AT yangliu exploringuavnarrowbandhyperspectralindicesandcropfunctionaltraitsderivedfromradiativetransfermodelstodetectwheatpowderymildew AT xiaoyangma exploringuavnarrowbandhyperspectralindicesandcropfunctionaltraitsderivedfromradiativetransfermodelstodetectwheatpowderymildew AT luluan exploringuavnarrowbandhyperspectralindicesandcropfunctionaltraitsderivedfromradiativetransfermodelstodetectwheatpowderymildew AT hongsun exploringuavnarrowbandhyperspectralindicesandcropfunctionaltraitsderivedfromradiativetransfermodelstodetectwheatpowderymildew AT fangkuizhao exploringuavnarrowbandhyperspectralindicesandcropfunctionaltraitsderivedfromradiativetransfermodelstodetectwheatpowderymildew AT xiaojingyan exploringuavnarrowbandhyperspectralindicesandcropfunctionaltraitsderivedfromradiativetransfermodelstodetectwheatpowderymildew AT yuntaoma exploringuavnarrowbandhyperspectralindicesandcropfunctionaltraitsderivedfromradiativetransfermodelstodetectwheatpowderymildew AT minzanli exploringuavnarrowbandhyperspectralindicesandcropfunctionaltraitsderivedfromradiativetransfermodelstodetectwheatpowderymildew |