Edges of target lead to noisy results

[roflmaostc]

If I optimize something with a hard edge, there is some noise on the boundaries.

Surprisingly the noise is over a region of 2 pixels and not only one.
Is there some inaccuracies in how the rays are traced?

In this case I had 32 rays per pixel, so some of them are gonna hit the object.

The noise is super annoying because it dims down all other pixel in our setup. So eliminating those high erroneous pixels is important.

files.zip

[roflmaostc]

[roflmaostc]

I think it is related to filter_radon=true. Without filter radon I get something like this.
Also something is a bit odd since the top line is slgithly too bright, especially in the hole. So as if it would try to polymerize this part even though it should not

config.json

[roflmaostc]

I think I understood what's going on.

If the sensor has a different pixel size than the projector, some of the rays from the same pixel will hit sometimes target voxels which are either inside or outside.

So to avoid those issues, the sensor pixel size should be an integer multiple of the projector pixel size.

[roflmaostc]

Here it can be seen. The written target intensity is in the wrong layer, because the corresponding pixel contributes to two different layers in the sensor discretization