-
-
Notifications
You must be signed in to change notification settings - Fork 246
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Image co-registration with AROSICS #148
Comments
hi @Noza1212 , this is a very good point. For Landsat images, the georeferencing accuracy is stored in the metadata (also in the |
Thank you, I'll look into it. |
yep, doco is here https://danschef.gitext-pages.gfz-potsdam.de/arosics/doc/ and there's also a paper on it https://www.mdpi.com/2072-4292/9/7/676 |
Hi @kvos I've tried to use arosics in order to correct the images. It's been quite useful so far, I've got great results with Sentinel 2 images. Nonetheless, I couldn`t get a result with Landsat images. I think it is because of the way Coastsat downloads Landsat images. |
sounds great, thanks for having a look! I can modify the way the images are downloaded if you let me know which bands/metadata is needed. Could you share a snippet of how you used the library to co-register the S2 images? |
Hi Kvos. I figured out the problem with Landsat images. It turns out that Landsat images come with missing data (-inf values) while Sentinel images do not. And with that, it should work just fine. One thing I noticed, though, was that with Landsat images the result of the co-registration process was not as good as the result I obtained with Sentinel 2 images. I think it's because Sentinel 2 images have a better resolution. |
hi @Noza1212 , thanks very much for reporting this, it is super helpful! I only got to read everything in detail now.
Please anyone who is interested in this feature feel free to participate to the discussion |
Hi Kvos. Sorry for the delay in answering to you. I've been kinda busy these days.
Finaly, I'm sorry if my english wasn't good enough. It's not my native language, so if I didn't explain myself good enough, please ask me again. I'm happy to help you with the enhacement of Coastsat, since it has been very useful to me. |
thanks for the reply, was very clear. Sounds good for this enhancement, hopefully a first version of this enhancement can developed this year on the |
Hi guys, My understanding is that we would need to mask the water pixels and shoreline + ~50m perimeter to avoid the image being corrected for shifts in the shoreline, is that correct? Cheers |
Hi everyone, I am also trying to incorporate Arosics into the CoastSat workflow. Real quick first, I just want to say how awesome this tool is Killian. In trying to use arosics, I'm running into a couple hurdles that it sounds like you all might be experiencing too. @Noza1212 , did you have any problem with CoastSat drawing a shoreline on the coregistered S2 image? I have successfully coregistered the 10m S2 tifs to a L8 image, but when I do the extraction step on CoastSat there is no shoreline drawn. The classification and MNDVI panels look good though. Did you experience this? |
hi @JanTiede and @shannonmgbrown , cool that you have been using AROSICS, that's really exciting. Sorry for the late reply from my end. |
@JanTiede , @shannonmgbrown , I created a new |
please check out the new |
I just want to give a small contribution to think on this issue. I found another approach useful for CoastSat. Andres Perez developed an ML tool for coregistration of a small time-series scenes: https://link.springer.com/article/10.1007/s00521-023-09242-0 repository is here: https://github.com/afperezm/multi-temporal-coregistration |
Dear Kvos.
I know Coastsat doesn´t correct the images and their correction is made by the data provider. But, I've noticed there´s an important gap or mismatch (up to 50 m in some images) between the images downloaded and an orthorectified image I got myself.
What would you suggest in that case? Is it ok if I rectify the images downloaded, delete the wrong ones and replace them?
Thank you in advance for your help.
Noza1212
The text was updated successfully, but these errors were encountered: