| 28 Jan 2014 02:41 AM |
|
|
Hi,
I am trying to atmospherically correct Landsat 8 data (OLI) using Flaash but I am getting negative values. I am first calibrating the image using envi’s “radiometric calibration” and applying flash settings (scale factor of 0.1) and then using a single scale factor of 1 for all bands in Flaash. This same settings work fine when correcting Landsat TM and ETM getting very accurate surface reflectance values but it doesn’t seem to work with Landsat OLI (even though I am using Envi 5.1 and Landsat-8 OLI is supported). Any idea why this may be? Thanks
Susana
|
|
|
|
MariM
Veteran Member
Posts:2396  
| 29 Jan 2014 12:02 PM |
|
|
There is an issue in FLAASH with Landsat-8 that requires you point to a different filter function in the Multispectral Settings. The default is for FLAASH to use the filter function from the \envi51\classic\filt_func directory but it needs to use the landsat8_oli.sli from the \envi51\resource\filterfuncs directory. Can you try setting the filter function to this newer location?
Some negative values in a FLAASH result is not uncommon, particularly in the visible bands. They tend to occur in dark areas such as shadow or water and are typically areas of low radiance that do not model well in FLAASH. It might be worthwhile to also test the data in Quac to compare the results.
|
|
|
|
Deleted User
New Member
Posts: 
| 03 Feb 2014 04:59 AM |
|
|
Hi MariM,
Thanks, pointing to the new filter function has solved the problem. I still have a few negative values but nothing to worry about. Many thanks
Susana
|
|
|
|