how to Convert MiDaS depth prediction to real-world distance?

[melaslalib]

0

I am using MiDaS (specifically the "DPT_Hybrid" from PyTorch Hub) to estimate depth from a single RGB image. MiDaS returns a depth image with values in the range 0-1.

I understand that these normalized values do not directly represent real-world distances. How can I convert the normalized MiDaS depth to meters or other real-world units?

my final goal is to get the real distance between the camera and the selected object from the depth image which i get it from monodepth using Midas.

I have tired to get the function between the this values and real world distance by measuring some know distance but didnt get results