If i use L8 and S2 scenes/granules as input in ACOLITE, which format does L8 and S2 have? DN, TOA reflectance/radiance...?
I got a bit confused, as few papers mentioned radiometric calibration from DN in reflectance and conversion in ACOLITE
Thank you
Input data
Re: Input data
Dear Anne
In short, the data is provided by the space agencies as top-of-atmosphere (TOA) reflectances.
The standard L1 files are integerised top-of-atmosphere (TOA) reflectances. Scaling and offset values are provided in the metadata to convert these integers to floating point values. ACOLITE does this rescaling automatically when processing.
As the TOA data are provided as scaled integers the data look similar to digital numbers (DN) but they are not strictly the count outputs from the analog to digital converter on board the satellite.
For Landsat 8 alternative scaling and offset values are provided to convert the integer numbers to TOA radiances. This option is available in ACOLITE (but not recommended), the TOA radiance is then rescaled to reflectance using the Thuillier irradiance dataset.
I hope this helps!
Quinten
In short, the data is provided by the space agencies as top-of-atmosphere (TOA) reflectances.
The standard L1 files are integerised top-of-atmosphere (TOA) reflectances. Scaling and offset values are provided in the metadata to convert these integers to floating point values. ACOLITE does this rescaling automatically when processing.
As the TOA data are provided as scaled integers the data look similar to digital numbers (DN) but they are not strictly the count outputs from the analog to digital converter on board the satellite.
For Landsat 8 alternative scaling and offset values are provided to convert the integer numbers to TOA radiances. This option is available in ACOLITE (but not recommended), the TOA radiance is then rescaled to reflectance using the Thuillier irradiance dataset.
I hope this helps!
Quinten
-
annetrampe
- Posts: 8
- Joined: Sun May 07, 2017 2:25 pm
Re: Input data
Thank you!
So the input is TOA reflectance (also for S2?).
In which step does ACOLITE rescale and for which purpose? Is this for the radiometric calibration?
For instance: I selected the Pahlevan L gains.
(I am currently trying to follow the calculations that are implemented in ACOLITE from TOA reflectance step by step until calculation of T and SPM)
So the input is TOA reflectance (also for S2?).
In which step does ACOLITE rescale and for which purpose? Is this for the radiometric calibration?
For instance: I selected the Pahlevan L gains.
(I am currently trying to follow the calculations that are implemented in ACOLITE from TOA reflectance step by step until calculation of T and SPM)
Re: Input data
Inputs are TOA reflectance! (In Vanhellemont & Ruddick 2014 we used Landsat radiances and converted those to reflectances.)
The gains in ACOLITE are applied on the TOA reflectances in order to mitigate the uncertainties in the sensor calibration and atmospheric correction. The gains provided are "system calibration gains" from processing with l2gen and not ACOLITE. (Note that even when you select the radiance 'L' gains they are applied to the reflectances unless you specify the radiance inputs for Landsat.)
The gains in ACOLITE are applied on the TOA reflectances in order to mitigate the uncertainties in the sensor calibration and atmospheric correction. The gains provided are "system calibration gains" from processing with l2gen and not ACOLITE. (Note that even when you select the radiance 'L' gains they are applied to the reflectances unless you specify the radiance inputs for Landsat.)
-
annetrampe
- Posts: 8
- Joined: Sun May 07, 2017 2:25 pm
Re: Input data
Is there any radiomatric calibration for Sentinel-2?
Last edited by annetrampe on Tue Dec 12, 2017 2:14 pm, edited 1 time in total.
Re: Input data
Currently only support for Landsat 8 is implemented. For other sensors this should follow in an upcoming version - possibly in the upcoming Python version.
-
annetrampe
- Posts: 8
- Joined: Sun May 07, 2017 2:25 pm
Re: Input data
Sorry, I think, I got a bit confused with the radiometric calibration:
So L8 and S2 are already provided radiometrically calibrated by the provider? (conversion of DN in TOA reflectance)
If I apply gains, is this some kind of calibration/correction?
I am sorry for the confusion. I am new to the field of remote sensing
So L8 and S2 are already provided radiometrically calibrated by the provider? (conversion of DN in TOA reflectance)
If I apply gains, is this some kind of calibration/correction?
I am sorry for the confusion. I am new to the field of remote sensing
Re: Input data
Hi Anne
The data provided by the space agencies are indeed calibrated TOA reflectances, not sensor counts or DN. The gains applied during the processing are an additional factor to compensate for a combination of remaining TOA calibration errors and the atmospheric correction system (in this case l2gen).
Quinten
The data provided by the space agencies are indeed calibrated TOA reflectances, not sensor counts or DN. The gains applied during the processing are an additional factor to compensate for a combination of remaining TOA calibration errors and the atmospheric correction system (in this case l2gen).
Quinten