A new remote sensing index of mildew for winter wheat disaster area extraction
Disasters caused by winter wheat mildew affect the quality and safety of grain. The purpose of this study was to construct a winter wheat mildew index based on multispectral remote sensing data combined with an automatic threshold segmentation algorithm to quickly extract the area affected by winter...
Uloženo v:
| Vydáno v: | International journal of digital earth Ročník 18; číslo 2 |
|---|---|
| Hlavní autoři: | , , , |
| Médium: | Journal Article |
| Jazyk: | angličtina |
| Vydáno: |
Taylor & Francis Group
31.12.2025
|
| Témata: | |
| ISSN: | 1753-8947, 1753-8955 |
| On-line přístup: | Získat plný text |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Shrnutí: | Disasters caused by winter wheat mildew affect the quality and safety of grain. The purpose of this study was to construct a winter wheat mildew index based on multispectral remote sensing data combined with an automatic threshold segmentation algorithm to quickly extract the area affected by winter wheat mildew. Research has shown that the spectrum of winter wheat strongly absorbs after mildew. On the basis of this spectral feature difference, the ‘area method’ is used to construct the wheat mildew index to enhance the separability between normal wheat and mildewed wheat. Five threshold segmentation algorithms are used to automatically extract the mildewed wheat region, and the Otsu threshold segmentation algorithm and iterative threshold segmentation algorithm achieve the best performance. The results show that the overall accuracy of the wheat mildew index combined with the automatic threshold segmentation algorithm is similar to that of the random forest and support vector machine (SVM) classification methods, but the extraction process is greatly simplified. In addition, the extraction of mildewed wheat combined with an automatic threshold segmentation algorithm can effectively avoid the problem of unstable extraction accuracy caused by specific threshold segmentations. |
|---|---|
| ISSN: | 1753-8947 1753-8955 |
| DOI: | 10.1080/17538947.2025.2553796 |