| 摘要: |
| 以地处南亚热带季风气候区的佛山市高明区山地为研究区，探索利用无人机获取高光谱影像
对森林可燃物类型进行划分的具体方法，将可燃物类型划分成针叶林、阔叶林、灌草地、竹林和非林地 5
个类型，采用无人机携带高光谱相机获取森林可燃物的高光谱影像设计针对森林可燃物的高光谱影像处
理流程并进行遥感影像处理和分类。结果表明通过有监督的机器学习分类方法可以实现对试验区主要可
燃物类型进行划分，经现场检验，分类准确度达到 81.94%。研究表明基于无人机高光谱遥感的森林可燃
物分类方法具备可行性。 |
| 关键词: 无人机 高光谱遥感 森林可燃物 遥感影像分类 |
| DOI： |
| 分类号: |
| 基金项目:广东省林业科技创新专项项目（2018KJCX003，2019KJCX011） |
|
| Study on Forest Fuel Classification Method Based on UAV Hyperspectral Remote Sensing |
|
Zhou Yufei, WANG Zhenshi, ZHONG Yingxia, LI Qiang, WU Zepeng, LI Xiaochuan
|
|
Guangdong Academy of Forestry
|
| Abstract: |
| Using the mountainous area of Gaoming district, Foshan city, which is located in the southern
subtropical monsoon climate zone as the research area, we explore the specific method of classifying the types
of forest combustibles using hyperspectral images acquired by UAVs, classifying the fuel types into five types:
coniferous forest, broad-leaved forest, scrub grass, bamboo forest and non-forested land. The results show that the
supervised classification method can be applied to the forest combustibles. The results show that the classification
of the main combustible material types in the test area can be achieved through the supervised machine learning
classification method, and the accuracy of the classification reached 81.94% after the field inspection. The study
shows that the method of forest combustibles classification based on UAV hyperspectral remote sensing is feasible. |
| Key words: UAV hyperspectral remote sensing forest fuel remote sensing image classification |
