Perfect Pixel overlap

Microwave Toolbox
#1

Hi everyone, I noticed that the pixels of different orbits from Sentinel 1 GRD data overlap perfectly. I guess that
this is caused by a resampling pre-process? If so, how can one generate GRD images which are not resampled in this manner?

Best regards,
Marvin

#2

by downloading the SLC data, I would guess :slight_smile:

#3

Okay, thank you! Which kind of processing steps do I need to perform with the SLC data in order to get GRD images, but not resampled in the way I mentioned above?

#4

please refer to the steps explained here:
http://sentinel1.s3.amazonaws.com/docs/S1TBX%20SAR%20Basics%20Tutorial.pdf

#5

Hi Marvin,

It can only be by coincidence that the lat/long of the same pixel coordinate from two S1 GRD products are the same. Due to slight orbit differences between successive orbit repeat periods (i.e. data acquired a multiple of 6 or 12 days apart) , it is expected that there will be slight differences in the pixel coordinate for a given lat/long.

Even if the pixel coordinates are the same, why is this a problem?

Peter

#6

Hi Peter,

what you mention is exactly what I expect: a slight offset due to slight orbit differences.
I have a research idea which requires the pixel not to overlap.
So far I figured I have to (1) download SLC data, (2) Apply precise orbits, (3) Deburst and (4) Finally perform mulit-looking.
This will hopefully result in images which have different pixel-sizes as well as different centroids for the pixels?

Best regards,
Marvin

#7

Hi Marvin,

So far I figured I have to (1) download SLC data, (2) Apply precise orbits, (3) Deburst and (4) Finally perform mulit-looking.

Your proposed steps will essentially create a GRD product and will require a lot of work to do want you want.

I have a research idea which requires the pixel not to overlap.
This will hopefully result in images which have different pixel-sizes as well as different centroids for the pixels?

In your first sentence you say you want pixels that do not overlap (i.e. have the same lat/long) but the second sentence you say the do overlap since they have different sizes and centroids.

For you idea have you considered the impact of speckle. If you have a uniform region which does not physically change between acquisitions, the individual pixels with the region will be different between acquisitions but the average pixel intensity will be the same (assuming there are enough pixels to calculate an accurate average).

Peter