4.3 Article

Spectral features and separability of alpine wetland grass species

Journal

SPECTROSCOPY LETTERS
Volume 50, Issue 5, Pages 245-256

Publisher

TAYLOR & FRANCIS INC
DOI: 10.1080/00387010.2016.1240088

Keywords

Alpine wetland grass species; spectral feature (characteristic) calculation; spectral separability; vegetation spectral feature extraction model

Categories

Funding

  1. Special Fund for the Ecological Assessment of the Three gorges Project [0001792015CB5005]
  2. National High Technology research and development program [2013AA12A302]
  3. Special Fund for Forest Scientific Research in the Public Welfare [201504323]

Ask authors/readers for more resources

This study aimed to analyze the spectral features of alpine wetland grasses and evaluate three types of spectral feature computational methods. In situ spectra of alpine wetland grasses were collected in the northeastern Qinghai-Tibetan Plateau (Niaodao Wetland of International Importance and Ruoergai Wetland of International Importance), and spectral features were extracted using the revised vegetation spectral feature extraction (VSFE) model, wavelet transform, and the characteristic band selection method based on trivariate mutual information. These methods were then compared for their ability to amplify spectral separability and reflect grass properties, using the Jeffries-Matusita distance and relative vegetation parameters, respectively. The results show distinct spectral features of alpine wetland grasses within the band from 717.80nm to 1050nm. These differences are chiefly caused by leaf area index and chlorophyll content, but not plant taxonomy. Among the three spectral feature analysis methods, the revised VSFE model provides the most information for separating different alpine wetland grasses, as well as indices reflecting chlorophyll, leaf area index, and nitrogen. However, only 70 alpine wetland grass species pairs among 138 can be clearly classified within the band from 350nm to 1050nm. Further study on wetland grass species classification over wider band ranges should be implemented. This study also provides suggestions for selecting a spectral feature computational method in hyperspectral data processing.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.3
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available