Hi there,
I am going to classify a scene and I am already stuck on the first step.
Could you tell me if the module ATMOSC can transfer the DN values directly into reflectance? or do I first need to convert it into radiance?
I am also wondering how to create a mask or how to eliminate cloud cover area from classification?
Regards,
Ilona
Ilona,
You are correct, ATMOSC will convert to reflectances.
As for clouds, one way is to just create a mask of cloud/shadow areas and mask your classification output.
James Toledano
Clark Labs
Thank you for your answer.
I have another question, does ATMOSC or IDRISI Maximum Likelihood classification deal with Landsat 8 data?
What about Tasselled cap model? is it just suitable for older Landsat 7 ETM+?
Regards,
Ilona
Yes, the module ATMOSC can be used to calibrate Landsat 8. You just need to get the parameters from the metadata. Once calibrated, you can certainly use Landsat 8 in the Maximum Likelihood routine.
In the current version, Selva, TASSCAP has not been updated with new coefficients for the transformation of Landsat 8.
James
sorry to be such a pain, however, I somehow can get it right. If you could help me and see what I am doing is fine or not.
I have chosen the Apparent Reflectance model, with Lmin and lmax.
For band 1:
Lmin RADIANCE_MINIMUM_BAND_1 = -63.65098
Lmax RADIANCE_MAXIMUM_BAND_2 = 789.28461
Max Value and display max = Dn max for this band: 38765.
However, in metadata stated quantised_cal_max_band_1=65535.
The result that I am getting is 1 throughout the whole image, for either Dn max.
However, if I divide the lmax and lmin by 10 (as suggested by help notes), I get values from 0 - 0.80, within the right range. Could you explain to me if that is right, what I am doing? Also, which DN max should I use?
Many thanks in advance,
Ilona
It would appear that what you are doing is correct. Each sensor can publish their data in different units, so unless you can get the details on the units used for the parameters, dividing by 10 or 100 or whatever, is what you will have to do until you get the correct results. That you got a result that is between 0 and 1 is a good sign that you now have the units correct for the Lmin and Lmax.
James