Stereo Vision Depth Triangulation

A robot with two eyes — a stereo camera rig — can recover distance the same way animals do: by comparing how far a point shifts between the left and right image. This simulator places a real left and right virtual camera on a movable rig looking into a cluttered 3D scene, renders both feeds live as inset views, and measures the actual pixel disparity of whatever object you select. That measurement is fed back through the triangulation equation Z = f·B/d to produce a depth estimate, shown alongside the true geometric distance so you can see triangulation succeed or fail in real time. Adjust the baseline and focal length to see the classic precision/field-of-view trade-off, and switch on pixel-quantization noise to watch depth error grow with the square of distance — the reason stereo rigs excel up close but lose accuracy far away.