artifacts with large area projections

28 views
Skip to first unread message

אליקום ודיסלבסקי

unread,
Nov 21, 2021, 8:07:35 AM11/21/21
to pyt...@googlegroups.com

Dear Pytroll group colleagues,

 

Hope all is O.K. with you.

 

We are at IMS are working on shifting all our python code to python3.

 

I've already made the necessary changes to use satpy with python 3.8 and it works fine,

However when I choose to work with very large area projections,

I get interesting artifacts("circles") at  the top of my images.

 

I've uploaded to google drive all the necessary data and code including images:

 

https://drive.google.com/file/d/1U56UK6Mxv9yGHeFBZfA7zaDwExAU8z4a/view?usp=sharing

 

Just download the data & code to a certain directory and run it from "Scripts" folder,

writing in the terminal: python3 multiplot_MSG_HRIT_RSS_2.py 201805131300

 

The output images are stored at "imageout" folder.

 

Your help for understanding the source for the artifacts will be appreciated.

 

Thank you,

 

Best regards,

 

Elyakom Vadislavsky

R&D department

Israel Meteorological Service

 

 

 

 

 

David Hoese

unread,
Nov 21, 2021, 2:11:32 PM11/21/21
to pyt...@googlegroups.com
Hi Elyakom,

I'm not a fan of downloading a random tarball, but even without looking
at your code I know what is likely going on. Try forcing
`radius_of_influence` when calling `Scene.resample`. You could try
`radius_of_influence=50000` first and see if that improves things.
Radius of influence tells the nearest neighbor resampling how far to
look for valid pixels before giving up. Satpy tries to guess the best
value to use, but in certain project combinations, especially those that
go to very high latitudes, it doesn't always do a great job. Note that
the higher this value the larger the "blurry" border is around the edge
of the image data.

Let me know if you have any other questions, but hopefully this gets you
a better result.

Dave
> --
> You received this message because you are subscribed to the Google
> Groups "pytroll" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to pytroll+u...@googlegroups.com
> <mailto:pytroll+u...@googlegroups.com>.
> To view this discussion on the web, visit
> https://groups.google.com/d/msgid/pytroll/b207901b6cd2447b9cbeb42cbe5ff49b%40MSMBX02.mot.dom
> <https://groups.google.com/d/msgid/pytroll/b207901b6cd2447b9cbeb42cbe5ff49b%40MSMBX02.mot.dom?utm_medium=email&utm_source=footer>.

אליקום ודיסלבסקי

unread,
Nov 22, 2021, 1:41:00 AM11/22/21
to pyt...@googlegroups.com
Dear Dave,

Thank you for your fast response and major help.

I've implemented your comment and it works fine, many thanks.

Attached the python code in PDF format to be shared with the forum.

Thank you,

Best regards,

Elyakom

To unsubscribe from this group and stop receiving emails from it, send an email to pytroll+u...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/pytroll/05327fb6-d877-5ef8-2a88-e337306d8fd6%40gmail.com.
multiplot_MSG_HRIT_RSS_2.py.pdf

Raspaud Martin

unread,
Nov 22, 2021, 5:22:34 AM11/22/21
to pyt...@googlegroups.com
Dear Elyakom,

Thanks a lot for the feedback and the code!

Best regards,
Martin
Reply all
Reply to author
Forward
0 new messages