Skip to main content

Accuracy and applicability of linear spectral unmixing in delineating potential erosion areas in tropical watersheds

Buy Article:

$63.00 plus tax (Refund Policy)

Abstract:

This study was undertaken to assess the accuracy of linear spectral unmixing (LSU) in estimating fractional abundance of land cover components and to examine its applicability in delineating potential erosion areas in tropical watershed. Five image end-members (mixed vegetation, grass, Acacia auriculiformis, bare soil and water/shadow) were selected and used in different combinations in unmixing Landsat Enhanced Thematic Mapper (ETM) into fraction images. The accuracy assessment was conducted by comparing the land cover abundance estimates derived from unmixing with the land cover abundance measured from field-validated classified QuickBird imagery. Good agreement was obtained using a four-end-member combination in which shadow was eliminated. The results suggest that LSU could be implemented for soil erosion detection. In general, soil erosion increases when vegetation cover decreases; hence, we used the fraction images to derive a bare soil/vegetation cover ratio and used that as a simple indicator to map high potential erosion areas. Comparison with field assessment of actual erosion levels in the study area showed that the technique is effective in identifying areas on which erosion control efforts should be concentrated.

Document Type: Research Article

DOI: https://doi.org/10.1080/01431160701874579

Affiliations: Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Japan

Publication date: 2008-01-01

More about this publication?
  • Access Key
  • Free content
  • Partial Free content
  • New content
  • Open access content
  • Partial Open access content
  • Subscribed content
  • Partial Subscribed content
  • Free trial content
Cookie Policy
X
Cookie Policy
Ingenta Connect website makes use of cookies so as to keep track of data that you have filled in. I am Happy with this Find out more