Continuum removal based hyperspectral characteristic analysis of leaves of different tree species
-
摘要: 高光谱遥感的出现使树种的精细识别成为可能,而高光谱数据具有波段多、数据量大、冗余度大等特点,利用高光谱遥感技术进行树种鉴别时,光谱特征的选择及提取是个非常重要的过程。选择了樟树Cinnamomum camphora,麻栎Quercus acutissima,马尾松Pinus massoniana和毛竹Phyllostachys pubescens 4个树种,利用包络线去除法对ASD高光谱仪实测的原始光谱数据处理,比较原始光谱和包络线去除曲线图,选择差异较大的波段用于识别不同树种,用欧氏距离法检验所选择的波段识别不同树种的效果。结果证明,利用波段较窄的高光谱数据能够挖掘出不同树种的光谱差异,实现不同树种的鉴别;包络线去除法能够有效解决高光谱数据冗余的问题,对4个树种叶片的高光谱进行波段选择,能够将有效波段减少到8个,为484 ~ 493,670 ~ 679,971 ~ 980,1 162 ~ 1 171,1 435 ~ 1 444,1 773 ~ 1 782,1 918 ~ 1 927和2455 ~ 2464 nm,并得到较理想的树种鉴别效果。图5表2参11Abstract: The development of hyperspectral remote sensing technology enables the precise identification of tree species possible. Because hyperspectral data are characterized by multiple bands,large database and great redundancy,the extraction and selection of the spectral characteristics is a very important process when identifying the tree species with hyperspectral remote sensing technology. The method of continuum removal was used to deal with the original spectral data of four tree species measured by ASD hyperspectral instrument. The curves of the original spectrum and the continuum removal were compared and the bands with greater differences were selected to identify the different tree species. Then the Euclidean distance method was used to test the selective bands identifying different tree species. The results showed that the continuum removal was an effective method of feature band extraction. Eight effective bands were selected with the continuum removal,including 484-493,670-679,971-980,1 162-1 171,1 435-1 444,1 773- 1 782,1 918-1 927 and 2 455-2 464 nm. They could identify different tree species effectively.[Ch,5 fig. 2 tab. 11 ref.]
-
Key words:
- forest management /
- tree species identification /
- hyperspectral /
- continuum removal /
- Euclidean distance
-
-
链接本文:
https://zlxb.zafu.edu.cn/article/doi/10.11833/j.issn.2095-0756.2010.06.001

计量
- 文章访问数: 6824
- HTML全文浏览量: 559
- PDF下载量: 4089
- 被引次数: 0