J. (Jochem) Verrelst, MSc: Space-borne spectrodirectional estimation of forest properties

  News
  Newsroom
  Archive
  Calendar
  2012
  2011
  2010
  2009
  2008
  2007
  2006
  2005
  2004
  2003
  2002
  2001
  2000
  1999
  News
  RSS
  Calendar
  Open days
  Courses
  Congresses and symposia
  PhD-graduations and speeches

7 Apr 2010 16:00
Unit: Wageningen University
Location: Aula, building 362, Gen. Foulkesweg 1, Wageningen
Organisation: Wageningen University
Promotor: prof.dr. M.E. Schaepman (Geo-information Science (Remote Sensing))
Co Promotor: Dr.ir. J.G.P.W. Clevers, Dr. B. Kötz (Application and Future Technology Department, European Space Agency-ESRIN, Italy)

 There is a significant increase in attention paid to the derivation of combined structural and biochemical properties of vegetation from space-based spectrodirectional observations. The CHRIS sensor onboard the PROBA satellite offers a combination of hyperspectral and multi-angular data from space. In my PhD research I estimated biochemical and structural properties of an Alpine forest using CHRIS data.

Firstly, the reflectance anisotropy of an Alpine forest was estimated by calculating a suite of broadband and narrowband vegetation indices from multi-angular data. The influence of both vegetation composition, as expressed by photosynthetic vegetation (PV) and non-photosynthetic vegetation (NPV), and structural variables on the anisotropic reflectance has been assessed. I studied how well chlorophyll content can be estimated from spectral data under varying forest conditions. Using the Rahman–Pinty–Verstraete model, the reflectance anisotropy has been quantified into a single parameter, the so-called Minnaert-k parameter. An analysis of the dynamics of the Minnaert-k parameter over the visible and NIR spectral region led to an improved understanding of the relationships between reflectance anisotropy and canopy structure. Finally, a mapping application was developed that quantifies forest cover heterogeneity at the subpixel scale by combining information from both the hyperspectral and angular domains.
Print this activity