Header

UZH-Logo

Maintenance Infos

Canopy biochemistry estimation using spectrodirectional CHRIS data


Huber, Silvia; Kneubühler, Mathias; Koetz, Benjamin; Schopfer, Jürg T; Zimmermann, Niklaus E; Itten, Klaus I (2006). Canopy biochemistry estimation using spectrodirectional CHRIS data. In: 2nd International Symposium on Recent Advances in Quantitative Remote Sensing (RAQRS), Torrent, Valencia, Spain, 25 September 2006 - 29 September 2006, 314-319.

Abstract

Sun and sensor geometry cause spectrodirectional effects in remotely sensed reflectance data which can influence the estimation of biophysical and biochemical variables. Previous studies indicated that vegetation indices can be strongly influenced by such effects and thus impact the results. This study examined the uncertainty induced by changing view angles on statistical methods used for nitrogen concentration (CN) estimation. We analyzed data of the spaceborne ESA-mission CHRIS (Compact High Resolution Imaging Spectrometer) on-board PROBA-1, which provides hyperspectral and multi-angular data with a spatial resolution of 17 m. The images were acquired in June 2005 over a test site in Switzerland and subsequently preprocessed. Linear regression models (LM’s) were developed between laboratory-measured CN, reflectance and transformed reflectance (continuum-removed and normalized), respectively, using a subset selection algorithm. For each CHRIS observation angle a particular LM was built. All LM’s were evaluated using 10-fold cross-validation with random splitting order of the data. By considering the adjusted R2 (adj.-R2), the root mean square error (RMSE) and percent error (% error), the LM’s were finally compared. Best CN predictions were achieved with models calibrated on nadir data with R2 of 0.63 and 0.59 using the reflectance and transformed reflectance, respectively. Generally, better LM’s were attained with nadir and -36° data than with +36° data (forward scatter direction), for both, reflectance and transformed reflectance. Applying nadircalibrated LM’s to off-nadir data was not successful to estimate CN. The results suggest that the CHRIS/PROBA mission provides useful data for biochemistry estimation. Caution is required when applying statistical methods developed on nadir data to data with directional effects.

Abstract

Sun and sensor geometry cause spectrodirectional effects in remotely sensed reflectance data which can influence the estimation of biophysical and biochemical variables. Previous studies indicated that vegetation indices can be strongly influenced by such effects and thus impact the results. This study examined the uncertainty induced by changing view angles on statistical methods used for nitrogen concentration (CN) estimation. We analyzed data of the spaceborne ESA-mission CHRIS (Compact High Resolution Imaging Spectrometer) on-board PROBA-1, which provides hyperspectral and multi-angular data with a spatial resolution of 17 m. The images were acquired in June 2005 over a test site in Switzerland and subsequently preprocessed. Linear regression models (LM’s) were developed between laboratory-measured CN, reflectance and transformed reflectance (continuum-removed and normalized), respectively, using a subset selection algorithm. For each CHRIS observation angle a particular LM was built. All LM’s were evaluated using 10-fold cross-validation with random splitting order of the data. By considering the adjusted R2 (adj.-R2), the root mean square error (RMSE) and percent error (% error), the LM’s were finally compared. Best CN predictions were achieved with models calibrated on nadir data with R2 of 0.63 and 0.59 using the reflectance and transformed reflectance, respectively. Generally, better LM’s were attained with nadir and -36° data than with +36° data (forward scatter direction), for both, reflectance and transformed reflectance. Applying nadircalibrated LM’s to off-nadir data was not successful to estimate CN. The results suggest that the CHRIS/PROBA mission provides useful data for biochemistry estimation. Caution is required when applying statistical methods developed on nadir data to data with directional effects.

Statistics

Altmetrics

Downloads

18 downloads since deposited on 30 Sep 2014
6 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Conference or Workshop Item (Paper), not refereed, original work
Communities & Collections:07 Faculty of Science > Institute of Geography
Dewey Decimal Classification:910 Geography & travel
Language:English
Event End Date:29 September 2006
Deposited On:30 Sep 2014 12:28
Last Modified:25 Oct 2016 21:13
ISBN:978-84-370-6533-5
Official URL:http://ipl.uv.es/raqrs/

Download

Preview Icon on Download
Preview
Content: Published Version
Language: English
Filetype: PDF
Size: 288kB