Radiometric & Geometric Correction Workflow


#101

Yes, I tried another images too, this is same problem.


#102

I have a used your suggested workflow for the GRD with topography.
The results are shown,after your proposed methodology, at the img below.

Can I extract amplitude images (VV-VH), variance texture images (VV-VH), contrast texture images (VV-VH) based on your methodology? And if so, at which step should I do that? And more important, how can I extract them?

Thank you in advance.


#103

The amplitude is the calibrated signal (Gamma0), nothing more to extract here.
Texture can be calculated with the GLCM module under optical > image analysis


#104

From the same image, depend what format to save and results difference

Pict 1: without Thermal Noise Remove, but save on Sigma dB to display

Pict2: without Thermal Noise Remove, save in complex, then save on SigmadB to display (so can see the lines)


#105

In the picture above is the result after the GLCM (Gamma VV Contrast). Here is the problem: The study area has NaN values, as well as other areas.
The image which I have download is Sentinel 1 IW GRD.
Before the GLCM step, I have apllied Noise calibration, Apply orbit file, Calibration, Radiometric terrain flattering & range doppler terrain correction.

Why I get these NaN values? How I can correct the problem?

Thank you.


#106

After which step do the NaN values occur?
Is there extreme topography in your image?


#107

After GLCM, at Gamma0 VV Contrast band. I dont have extreme topography (Study area Lesvos island, Greece)


#108

if the NaN occurs after GLCM it means that the texture is 0 at those locations. You can disable the NaN display in the band properties by unckecking “No Data Value Used”


#109

Do you find out what the problems and how to resolve? Thanks advantage.


#110

I have no idea, sorry


#111

Somebody already mention about 2 years ago


#112

Problem I believe about save from ThermalNoiseRemoval, if s/w continue with complex they can able to save
complex and the lines on images will be solve too.