Scene width and height of sentinel-1

My main aim is reprojection and resampling.

First I reprojected sentinel-1 image from WGS84(DD) to UTM Zone 18 and then I resampled to 1000m. Once the image is resampled I reprojected back to WGS84(DD) since i prefer to work in Lon and Lat.

Now, my question is why I see big difference in scene width and height between resampled image in UTM zone 18 (285x213) and WGS84 (DD) (367x214). According to my unserstanding there might be differences in scene width and height but why it so big specially for the width?

The second question is why SNAP gives the coordinates of UTM in degrees while it should be in meters?

why don’t you just resample in WGS84 with the 1000m given in degrees?
I have the feeling that lots of information is unnecessarily lost during the reprojection to UTM12 and back.

because resampling in degrees is not possible. That is why I converted to UTM. Is is possible in degrees?

you can calculate the distance of 1000m in degrees according to the latitude of your study area.

In resampling parameters page there are three possible ways:

  1. By reference band from source product
  2. By target width and height
  3. By pixel resolution (in m)

SO I use the third option. In this option, only meter is shown in the bracket.

Based on your suggestion I tried,but it doesnot work. It oly gave me 1X1 pixel.

And what happens when you enter 1000 m for the WGS84 projected data?

I recommend bilinear downsampling here so your new pixels are an average of the 10m ones.

I want downsampling by mean from 10m to 1000m. It works for UTM but not for wgs84. Because UTM is in meters. But when I enter 1000m by changing to degree it gives me only 1X1 pixel.

you are right, the pixel dimensions are not given in meters in the third resampling option when the data is projected in WGS84. @marpet is this a bug?

How about using the multi-looking for this? Works before terrain correction, but at least it does what you want. .


I think this is a good approach for my intent. Thank you very much!

Yes, it is an issue. Wee have already a ticket for it.
The multilook is a good alternative for the S1 data.
For others it is possible to calculate the expected width or height and specify this value.
In any case this should be fixed. I’ve increased the priority for the issue.

1 Like