What accuracy can openCV provide with depth map?

Question

Python

I want to organize stereo vision using OpenCV and I would like to know what positioning accuracy can be obtained in this way? I perfectly understand that this largely depends on the cameras, but I would like to get an accuracy of 0.1mm at a distance of 0.5 meters. Is it possible?

Answer 1

No. We need to look for additional means of clarification

Answer 2

In order for the accuracy to be 0.1mm by 0.5m (500mm), the resolution of the image (vertically most likely) representing the depth map must be at least 5000 pixels (500 / 0.1 = 5000). I think if the resolution is 4-5 times higher than this value, then it will be possible to achieve close accuracy (but the speed of building a depth map for such a large image will be very slow). So it really depends a lot on the cameras.
I made such a camera, the main point was the calibration of the cameras (and obtaining a distortion matrix) and the exact relative position.
In the version you described, it is difficult for me to imagine the shooting conditions.