After radiometric correction, the data ignore value is set in the header and appears to be causing Quac to fail. Try the following: - select the file in the Layer manager, right click and choose View Metadata - in the Metadata Viewer, select Edit Metadata - in Edit Metadata, select Remove and find the Data Ignore value. Remove this value. - select OK to close the dialog. The file should be closed and re-opened with the new information. Try running Quac on this result.

Hello everyone! I' ve applied QUAC, as well as FLAASH method, so as to correct my image. I' ve applied both the methods to 198 bands, the radiometrically corrected. The results from QUAC include a small number of pixels with values greater than 10,000. As regards to the FLAASH method, the results include pixels with reflectance values lower than zero. So, I can't understand which method is better. What's the allowed percentage of pixels with negative values for the FLAASH method? Furthermore, teh bands with the negative values of pixels might be excluded because i don't remove the bad bands yet! Thank you very much for your help!

There really is no way to tell which one is more "correct" unless you have a reference image. These are two different methods (FLAASH and Quac) that estimate surface reflectance in very different ways. FLAASH is based more on physical models and is much more sensitive to the input parameters used while QUac uses in-scene statistics to estimate surface reflectance and does not rely in user supplied parameters. Both methods were developed by the same scientists as Spectral Sciences Inc and are typically similar but not exactly the same.
There are no specified number of 'allowed' pixels that go negative. If you have a lot of shadow or water in your image, you are likely to have more negative pixels. If you have a lot of bright cloud or corner reflectors, you could have some values that exceed 10000 (100%) because those pixels do not model well. Both FLAASH and Quac detect bad bands and set them in the output image header file so you should already have the 'bad bands' as determined by the algorithm(s).

That was very heplful! Thank you very much!!! About the bad bands you mean those that have been set with a zero at the output file?

The bad bands will be set in the image header (.hdr) file and you can view which bands are set to bad in the View Metadata tool. Also, if you bring up a spectral profile, you will see missing sections of the spectrum where the bands are set to bad bands.