| 19 Oct 2011 09:24 AM |
|
|
Hi everyone! I´m trying to apply an atmospheric routine with FLAASH (ENVI 4.5) in a Quickbird image. The scene has just benthic substrate (covered by water) and a few small islands. Because of this, I´m using the Maritime Aerosol Model. I have used different "Visibility Initial" values, but in all cases the values are very low. Even, in the best result, red and IR bands have almost all pixels with negative values. Does anyone know the reason? Any advice?
Thank you!
Laura
|
|
|
|
MariM
Veteran Member
Posts:2396  
| 20 Oct 2011 07:20 AM |
|
|
FLAASH is not really designed to correct over water images. Did you convert your QB image to radiance? What did you use for the scale factor? Incorrect assumptions about the data units and an incorrect scale factor are the most common causes of a predominance of negative reflectance values. However, water and shadow areas with low radiance do not model well and can also return negative values.
|
|
|
|
Deleted User
New Member
Posts: 
| 28 Oct 2011 10:09 AM |
|
|
Hi MariM, Thanks for your answer. I converted DN to radiance using floating data (no factor scale). When I applied FLAASH, I used factor scale 1. Also, FLAASH failed to correct WorldView-2 images with similar targets.
Regards,
Laura
|
|
|
|
MariM
Veteran Member
Posts:2396
| 28 Oct 2011 12:04 PM |
|
|
How did you calibrate your data? The standard coefficients convert IKONOS to W/m^2*sr*um which would require a scale factor of 10 in FLAASH to get to units of um/cm^2*sr*nm. ENVi 4.5 does not have a WV-2 calibration tool so you would also have to convert this to appropriate units as well. In what way did it 'fail to correct Worldview-2 images with similar targets'? It might be best if you contact technical support directly.
|
|
|
|