I would like to ask a question regarding the usage of C2RCC algorithm for Landsat 8 data and more specifically if there is a way to mask or extract the water body since the result contains the whole area of interest (not only the water body but land areas as well). I had no similar problems applying the algorithm on Sentinel-2 data. I have included an image so it will be clear as to what I mean. I am interested only in the water body (area marked with black).
Thank you very much in advance for your help .
Unfortunately, the Landsat data does not provide a land-water-mask.
To solve this problem, you can use some pixel identification too upfront and at the result to your data. Then you can use it in the valid pixel expression of C2RCC.
Please see theses posts.
You could also use Idepix for Landsat 8.
For S2 MSI such expression on the radiances is defined. But not for L8. I will check why and with the next updated we might provide a default expression.
One reason that no expression is provided is that L8 does not provide bands at an appropriate wavelength to discriminate properly between water and land.
Another option to create a usable mask is the fractional land-water mask or the land-sea mask operator both located in the menu
Raster / Masks
actually there is an appropriate band available.
The near_infrared (B5) can be used.
The problem is that the user can select if the values in the product shall be shown as radiances or reflectances. So, fixed values cannot be given here.
By default the values are radiances. In this case you can use this expression:
near_infrared > 0 && near_infrared < 25.
If you have switched to reflectances, the same values as for S2 should work.
near_infrared > 0 && near_infrared < 0.1
Maybe you need to change the values a bit.
Thank you very much for the different solutions. I will try them all and i will come back to let you know which one worked better for this case.