A stereo sensor is composed of two cameras to observe the scene. The orientation of the sensor in the camera is represented by two rectangles. By finding corresponding points in two images, constructing their rays, and intersecting them in space, a point is reconstructed in three-dimensional space. Note that the angle between the cameras is exaggerated to show the geometry of the stereo sensor more clearly.
The principle of the stereo sensor is the same as that of the HVS: two cameras are used to observe the scene from different positions. In order to correctly reconstruct the moving objects, the synchronization of the cameras is crucial.
Different viewpoints of the camera result in different positions of the points at a certain distance from the camera in the two images, which creates parallax. The parallax is the basic information to realize the three-dimensional reconstruction of the points in the scene.
