I have two satellite images, one for LandsatOLI and the other Sentinel 2, I have done the process , I calculated NDVI for each image (after resampling OLI), but I find that is a difference between the two NDVIs, so I’m looking for how I would adjust NDVI of OLI to be similar to that of sentinel2
I read the literature and I find that there a coefficients for the intercalibration between the different sensors
Then I’m looking for is what there are coefficients for intercalibration of NDVI between LandsatOLI and Sentinel2
I stumbled upon an interesting comparison just yesterday, maybe you haven’t seen it yet. It is not about calibration but closely looks at the differences of both sensors. www.mdpi.com/2072-4292/8/12/1014/pdf
Besides that, I am also interested in the intercalibration issue.
I did the resampling for a Landsat OLI image at 10m, after I calculated NDVI, and I compared this result with NDVI calculating from a sentinel2 image , there is much difference.
This is why I want to do intercalibration.
I don’t think that resampling is a good way because the information contained in different spatial resolutions can not be reconstructed by changing the raster size. It will be always a mixed signal of different surface emmissions.
Did you run sen2cor for Sentinel-2 and the calibration of Landsat (with SCP, for example)?
Yes I did the treatment with Sen2cor for the sentinel image, Landsat image was processed with SCP.
I want to build a time series to do the mapping of different cultures, but I find sentinel images with clouds in our study area, which is why I also want to use Landsat images.
But I do not know how to solve this problem of intecalibration.
I have done the same with both sensors and found the same problem. First of all: ¿Which bands are you using in SENTINEL? In my opinion choosing 8A for the infrarred is more accurate altough we loose spatial resolution( from 10 to 20 meters). Center and wide of this band is almost equal than LANDSAT8 5 band.