Open Collections

Radiometric correction of satellite imagery for topographic and atmospheric effects

University of British Columbia

The radiometry of satellite imagery is influenced by ground cover, local topography, and atmosphere. In order to increase the accuracy of ground cover identification from satellite imagery, effects due to topography and atmosphere must be removed. These effects can be estimated by modeling the image-formation process. For this thesis an image-formation model is developed and tested on Landsat MSS data over a mountainous region. Solar illumination angle, atmosphere depth, and sky illumination are calculated with the help of a digital elevation model. A digital forest cover map is used to select a target forest type for which model parameters are estimated using regression analysis. Results of this analysis indicate that solar illumination angle has the largest effect on target pixel irradiance followed by atmosphere depth. Sky illumination as calculated, was significantly correlated with target pixel irradiance but in a negative sense. This correlation suggests that inter reflection (also called mutual illumination) from adjacent terrain may be a small but significant source of illumination. The estimated model parameters are used to correct the imagery for topographic and atmospheric effects. Visual assessment of the corrected imagery indicates that many but not all of the topographic effects have been reduced. Comparisons between computer classified imagery and the forest cover map show an improvement in correctly classified pixels from 54% for the original image to 72% for the corrected image.

Text

2010-06-20

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/25890

Forestry

University of British Columbia

Graduate