I am trying to extract the depth of the unreal environment for a machine learning application, but I cannot seem to get it to work properly.
I have tried using a camera set to “SceneDepth in R” but I cannot easily extract the results. If I use a render target and then export it as an HDR file, the pixel values no longer correspond to the actual in-game distances. I have also tried to make a material, and then a texture, from the render target and export that, but that also didn’t preserve the values.
I have also tried using render passes (using scene depth) in the sequencer but this also didn’t work.
Is there some way to export the SceneDepth in R capture to a CSV or similar non-image format so that the actual in game distances are preserved?
I am still fairly new to unreal so simpler explanations or links to tutorials would be greatly appreciated.