Constraints on geometry
Because what I want to adjust in the end is still a binocular camera, the most direct way to find the correspondence between the two images captured by the two cameras is to match point by point, which is very cumbersome. Constraints For polar constraints, the search range can be greatly reduced.
Premise: Cameras C1 and C2 are on the same straight line: the two cameras are coplanar and their optical axes are parallel, and the parameters are the same (because I want to do dual purpose in the end)
Baseline [baseline]: The straight line C1C2 is the baseline.
Epipolar Plane: Any plane that contains the baseline is called an antipolar plane.
Epipolar line: the intersection line between the epipolar plane and the image.
Epipole: The intersection point of the camera's baseline with each image
Epipolar Constraint: Correspondence between points on two polar lines.
C1, C2 are two cameras, P is a point in space, P and the two camera center points C1, C2 form a plane PC1C2 in three-dimensional space, called the polar plane. The polar plane and the two images intersect in two straight lines, called polar lines. The imaging point of P in camera C1 is p1, and the imaging point of P in camera C2 is p2. For point p1 in the left picture, find its corresponding point p2 in the right picture, so that the spatial position of point P can be determined, which is the distance (depth) between the space object and the camera we want.
The epipolar constraint means that when the same spatial point is imaged on two images separately, and the projection point p1 of the left image is known, then the corresponding projection point p2 of the right image must be on the epipolar line relative to p1, which can be greatly reduced For the matching range, you only need to find p2 on the level line, not on the whole picture.
