(How) is the rasterio.merge function taking into account different crs for the rasters

[VehpuS]

I have been looking into the implementation of rasterio.merge.merge to "prefetch" common bounds from a group of rasters. Once working on my implementation (since the function isn't currently too modular) I realized that I was getting errors when calculating common bounds of tiffs.

After some investigation, it turns out this was because the CRS of each tif was different, resulting in bounds given in different units. Before implementing my own solution, I looked at the merge implementation again but found no consideration for this. In the end, I implemented a solution based on rasterio.warp.transform_bounds.

I'm writing here to check if I missed something. If not, I would be happy to make changes to the merge function to either fix this or specify the assumption that all tiffs (and the target bounds parameters) are assumed to share the same CRS before running merge. I have a good approach for a solution if necessary, but I wanted to confirm my suspicions before implementing a PR.

[VehpuS]

Any chance to get a response? I may try to send a pull request with an additional merge implementation that takes CRS into account to see if it's useful, get the conversation going, but I figured a prior discussion may be better :) .

[snowman2]

rioxarray wraps rasterio merge and it re-projects the rasters to the same CRS and resolution before merge. Here is a link to the code ref. It may be a helpful reference.

[snowman2]

You may be interested in #2573 if your main goal is to get the bounds and you have rasters of different CRS.

UPDATE: After further inspection, that is not the case. They must have the same CRS.

[sgillies]

Merging is requires all rasters to have the same CRS and is optimized for this case.

[groutr]

Merging is requires all rasters to have the same CRS and is optimized for this case.

Perhaps merge could verify that CRS is same across inputs and issue a warning if there is a difference? Merge reads metadata from first raster, it could cache CRS. Then during merge, when looping thru all the datasets, compare each CRS with first raster's CRS. Issue a warning if CRS doesn't match. That should have a minimal impact on performance.

[VehpuS]

Merging is requires all rasters to have the same CRS and is optimized for this case.

Perhaps merge could verify that CRS is same across inputs and issue a warning if there is a difference? Merge reads metadata from first raster, it could cache CRS. Then during merge, when looping thru all the datasets, compare each CRS with first raster's CRS. Issue a warning if CRS doesn't match. That should have a minimal impact on performance.

If that's acceptable, I'd love to create a pull request for this. While the docs mention the CRS having to match I think it's always better to provide active warnings (it took me personally a while to realize this was even an issue).

[sgillies]

Sounds good @VehpuS ! It would be much appreciated.