The output color of land in the 3 adjacent scenes in sentinel 1 GRD after preprocessing is different

Dear Step Forum Users,
I have a question regarding the output color of 3 adjacent scenes of sentinel 1A-IW_-GRDH images.
I have done pre-processing steps for the three images as follow:
1- apply orbit to file
2- thermal noise removal
3- radiometric calibration (singma0)
4- speckle filter
5- terrain correction
as a result, the output color of landcover in 1 image is different from other two, I don’t know why. I think they should be the same color when i use “open RGB image window”. pls see the output color below
My questions are:
1- The order of these steps for pre-processing is alright or not?
2- Why the output color of land features in one image is different from others when i use “open RGB image window”
3- How to process 3 images with the same color display on land cover?
Thanks,
Thoa

Probably this happen because SNAP cut’s off the the lower 1% and the upper 4% of the histogram to get rid of outliers.
You can set the histogram stretch to 100% for each channel, then they should look the same.
Or you can change the RGB profile. I made the following expressions for Sentinel-2 images, but you can adapt them to Sentinel-1.

blue=B2 < 0 ? 0 : B2 > 0.25 ? 0.25 : B2
green=B3 < 0 ? 0 : B3 > 0.3 ? 0.3 : B3
red=B4 < 0 ? 0 : B4 > 0.35 ? 0.35 : B4

The expressions remove all values below zero and above a certain value. You need to find good thresholds for you data.

1 Like

Thanks Marpet,
I have solved the problem as per your suggestion. Actually, I just simply subset and then process it. it works.
Thoa

A post was split to a new topic: TOPSAR-Merge: Source product should not be map projected