Image and Signal Processing Methods

Estimating wetland vegetation abundance from Landsat-8 operational land imager imagery: a comparison between linear spectral mixture analysis and multinomial logit modeling methods

[+] Author Affiliations
Min Zhang, Wenji Zhao

Capital Normal University, College of Resource Environment and Tourism, Beijing 100048, China

Zhaoning Gong

Capital Normal University, College of Resource Environment and Tourism, Beijing 100048, China

University of South Florida, School of Geosciences, 4202 E, Fowler Avenue, NES 107, Tampa, Florida 33620, United States

Ruiliang Pu

University of South Florida, School of Geosciences, 4202 E, Fowler Avenue, NES 107, Tampa, Florida 33620, United States

Ke Liu

National Administration of Surveying, Mapping and Geoinformation, Satellite Surveying and Mapping Application Center, Beijing 101300, China

J. Appl. Remote Sens. 10(1), 015005 (Jan 20, 2016). doi:10.1117/1.JRS.10.015005
History: Received July 3, 2015; Accepted December 10, 2015
Text Size: A A A

Abstract.  Mapping vegetation abundance by using remote sensing data is an efficient means for detecting changes of an eco-environment. With Landsat-8 operational land imager (OLI) imagery acquired on July 31, 2013, both linear spectral mixture analysis (LSMA) and multinomial logit model (MNLM) methods were applied to estimate and assess the vegetation abundance in the Wild Duck Lake Wetland in Beijing, China. To improve mapping vegetation abundance and increase the number of endmembers in spectral mixture analysis, normalized difference vegetation index was extracted from OLI imagery along with the seven reflective bands of OLI data for estimating the vegetation abundance. Five endmembers were selected, which include terrestrial plants, aquatic plants, bare soil, high albedo, and low albedo. The vegetation abundance mapping results from Landsat OLI data were finally evaluated by utilizing a WorldView-2 multispectral imagery. Similar spatial patterns of vegetation abundance produced by both fully constrained LSMA algorithm and MNLM methods were observed: higher vegetation abundance levels were distributed in agricultural and riparian areas while lower levels in urban/built-up areas. The experimental results also indicate that the MNLM model outperformed the LSMA algorithm with smaller root mean square error (0.0152 versus 0.0252) and higher coefficient of determination (0.7856 versus 0.7214) as the MNLM model could handle the nonlinear reflection phenomenon better than the LSMA with mixed pixels.

Figures in this Article
© 2016 Society of Photo-Optical Instrumentation Engineers

Citation

Min Zhang ; Zhaoning Gong ; Wenji Zhao ; Ruiliang Pu and Ke Liu
"Estimating wetland vegetation abundance from Landsat-8 operational land imager imagery: a comparison between linear spectral mixture analysis and multinomial logit modeling methods", J. Appl. Remote Sens. 10(1), 015005 (Jan 20, 2016). ; http://dx.doi.org/10.1117/1.JRS.10.015005


Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Related Book Chapters

Topic Collections

Advertisement
  • Don't have an account?
  • Subscribe to the SPIE Digital Library
  • Create a FREE account to sign up for Digital Library content alerts and gain access to institutional subscriptions remotely.
Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).
Access This Proceeding
Sign in or Create a personal account to Buy this article ($15 for members, $18 for non-members).
Access This Chapter

Access to SPIE eBooks is limited to subscribing institutions and is not available as part of a personal subscription. Print or electronic versions of individual SPIE books may be purchased via SPIE.org.