Error when running orthority using multiprocessing

[fakurten94]

Hello,

I have been using orthority and saw that there's a recurring error when running multiprocessing. This does not stop the calculation, but it starts spamming the error. Here's a snapshot of what I mean:

The function orthorectification is the following:

def orthorectification(file,out_fld,dem_file,cameras):

    camera = cameras.get(file)

    ortho = oty.Ortho(file, dem_file, camera=camera, crs=cameras.crs)
    ortho.process(f'{out_fld}/{os.path.basename(file)}',compress="deflate",interp="bilinear",overwrite=True)

Not sure if I could use orthority using Multiprocessing as it seems that it is using a good amount of memory trying to orthorectify a single image. I also have used the same process in a for loop and does work with no issues, this just becomes a problem when running it through multiprocessing. Any advice on how to mitigate that issue?

For reference, I have been using a Mac with Intel chip.

Thanks!

[dugalh]

Hello

Thanks for the report. It looks like the same issue as this. Orthority uses the PROJ library via Rasterio and configures it for on-demand caching of transformation grids.

I will look into this. You could try downloading the transformation grids for your area and see if that helps.

Bear in mind that Orthority is already multi-threaded so you are unlikely to see much speed-up from multi-processing. And yes, memory could be a problem.