running s2p on image crops

[A-Jacobson]

I'm trying to run s2p on worldview3 pairs without downloading the entire 40k by 40k full-images. I can do this with nasa ASP by placing my crops of interest in their appropriate locations within an empty numpy array the size of the full-image and supplying a --left-image-crop and a --right-image-crop at runtime. Is there a way to manually set a crop for each image to s2p? If not, what would be the best way to run s2p on image crops without ever downloading the full images?

[carlodef]

You can run s2p without downloading the input images by prefixing the url of your input images with /vsicurl/. Support for GDAL vsicurl was introduced in pull request #145

[A-Jacobson]

Hi Carlo, thanks for your reply. I've been learning a lot going through your library.

Unfortunately, I'm not aware of a way to get a url of the full image in a straightforward way. My understanding is that digital globe stores full images in dask arrays distributed across many machines and that compiling a full image's dask array to a geotiff takes a lot of bandwidth and time. Using digital globe's api I am able to access geotiffs for crops of the full image in a cheap/quick manner. My areas of interest are usually the size of a few houses or less than 500 by 500 pixels so I am hoping to find a way to run the pipeline using just those image crops.

The problem with the cropping approach has been that I do not know how to crop rpcs and haven't been able to find any literature on the topic. I saw your rpc_cropper tool but that tool takes a full image as input as well. Is there a way to get this to work with s2p? or perhaps a way for me to crop the rpcs on my own without using the full image as input?

[BryceCicada]

Hi,
This is exactly what I'd like to do. @A-Jacobson, did you get anything like this working?