I’m trying to figure out a way to develop a 2d-3d correspondence for stereo camera calibration. One idea I had was to “group” points together based on their approximate radial line/central angle. Some visualized examples are attached below. Unfortunately, they’re not perfect because I slapped everything together rather quickly for the sake of a quick experiment, but you get the gist
Left image + some lines and points:
Right image + some lines and points:
So I assumed that all the radial lines, denoted by the green lines and the red/purple lines that are reasonably close to those green lines, all converge to the camera center coordinates. Is this a valid assumption?