Hi,
I have started exploring preprocessing Sentinel 1 data, and have found lots of useful threads here already.
However, I have found nothing on when users clip their graticule to an area of interest to save on processing power and time.
When is the earliest point this can be done?
Many Thanks
Will
You can SubSet or Collocate or use Mask(s). Native functions in the interface, callable by GPT and/or snappy. So basically you should be more than fine with these options.
For Sentinel-1 there are other options for Interferometric stacks.
Cheers
Petr
Thanks for the reply Petr, where can I find these options?
Sorry I’m unfamiliar with what GPT and snappy are
Cheers
Will
Snappy is Python binding to SNAP Toolbox or rather the other way around and GPT is command line tool that comes with SNAP Toolbox basically it is the Toolbox without UI.
In the regular UI these options can be found under the “Raster” Submenu.
Hello @ABraun and @marpet
I have an issue with clipping my processed sentinel1 image. I will be using this data for crop classification and I have a polygonal shape file for the AOI.
I used the land/sea mask and got a weird result. I am attaching the screenshots below, can you please look into it?
1st image shows my polygonal shape file.
2nd image shows the parameters
3rd image shows result1
4th image, i have selected the option “invert vector” incase it changes my result
5th image shows the same result as the result1
I even exported the SAR image into tiff in QGIS and the file size is 7 GB, is that normal? I am using GRD format. Also I tried clip raster by mask layer in that but it shows the following error.
Can anyone please look into this issue as sooner as possible?
Image 3 looks like the result you are looking for. Why do you want to invert it?
I tried the inversion option and it worked for me.
Regarding the files size. Your scene is quite huge. Depending on the data type of the bands this might be.
You could try NetCDF4-CF as data format. This might compress the data better.
For such big geotiff files compression should be enabled by default too.
Maybe something went wrong.
Could you send your log file, if you still have it?
See here on [how to find the log files].(https://senbox.atlassian.net/wiki/spaces/SNAP/pages/1898053693/SNAP+FAQs#How-to-find-the-log-files?)
Hey Marpet,
Thanks for the prompt response.
I do not have the log files now. But, this file which is loaded in the SNAP is in the format of BEAM-DIMAP.
I tried the inversion option incase if there is sth wrong with it.
What shall I do? Shall I export it as Netcdf4 and then import it in Netcdf4 and later apply masking on that?
If that doesn’t work either, do you know how can I load Beam-Dimap into QGIS as Geotiff? I tried it but the clipping raster by mask layer gives the above error.
Thanks a lot.
If you have stored the product as BEAM-DIMAP you find envi files in the *.data directory.
I think you can load them into QGIS.
Hey Marpet,
I found the messages log file.
Please find it attached in here.
messages.log (79.1 KB)
Thanks.
ENVI files are into .hdr format. I haven’t worked with those previously, are they similar to tiff and are they as big as geotiff?
Thanks.
Yes, they are also quite big. Because they are uncompressed. But there is one file for each band which makes them smaller on the other hand.
Thanks Marpet. I think that will help now.
Hey @marpet and @ABraun,
I have one more question.
I want to undergo crop classification and for this I am following multitemporal method for 5 months.
Now my area of interest has three SAR images, which should be mosaicked at the end and then clipped to get the final ROI.
I am facing issue with coregistration. Two of the SAR images has the same orbit path but the other third image has different orbit path.
So I followed the preprocessing steps individually for all of them since batch processing takes more than usual time on my system. I followed the following process:
after i performed these steps individually for every image, I want to coregister them. But I have two different orbit paths. Yet, I tried the coregistration method and it was successful but I did not get proper result. Most of the resultant slave images are not present.
In short, I want a multitemporal file for my ROI which needs three images of SAR to be merged.
How can I correct my steps? I want three polarizations in the db format (VV, VH and VH-VV) to get a multitemporal stack at the end. I even have to mosaic and then clip these multitemporal images at the end to get my ROI. I am confused a lot.
I am grateful if someone could answer me as early as possible.
Thank you so much
You can only coregister images of the same orbit in SNAP. If the geometric correction with RD Terrain Correction worked well you could simply stack all images using the Create Stack operator or with the Collocation tool.
But my images have different orbit paths. Will create stack work for such images? I am confused about that. Also, can you explain me what “multitemporal compositing” is in Radar? I saw one of the video online and the person applied this after coregistration and this (“multitemporal compositing”) combined all the bands into single VV and VH band. Does this undergo loosing the metadata or affect the backscatter intensities?
Thank you ABraun for the quick response.
Also, I have to mosaic and clip to my ROI. How can I do this after creating stack?
