Anyway to speed up orthophoto merge in split merge step
I’ll leave that to Piero, but that step is pretty highly optimized. There’s of cool stuff under the hood to make that run as fast as it does.
The algorithm is pretty efficient, I can’t think of a lot of ways in which it could be sped up (doesn’t mean it’s not possible) without sacrificing the memory bounds guarantees that it offers (you can merge very large orthophotos without a lot of memory). You could try to patch the code to use larger window sizes for read/write and see if that improves performance?