Method for calibrating parabolic catadioptric camera using spatial straight lines

Abstract

The invention relates to a method for calibrating a parabolic catadioptric camera using spatial straight lines. The camera is calibrated using three straight lines in space as targets. First, in a unit virtual globe model, three non-parallel lines in space are projected and are projected onto an image plane to form three intersecting quadratic curves, and because the two quadratic curves intersectat two real intersecting points, the two intersecting points form a straight line. However, due to the presence of noise, the three straight lines do not intersect at one point but form a triangle. The center of gravity of the triangle can be considered as the approximate principal point on the image plane. Secondly, epipolar lines of the principal point about the three quadratic curves are solved respectively; as can be learnt from the relationship between the epipolar lines of poles, the epipolar line of the principal point with respect to the quadratic curve is a vanishing line, and each vanishing line and the corresponding quadratic curve are intersected at an image of two ring points, and thus intrinsic parameters of the camera are solved.

Description

technical field [0001] The invention belongs to the field of computer vision and relates to a method for solving internal parameters of a parabolic catadioptric camera by using three straight lines in space. Background technique [0002] The purpose of computer vision research is to allow computers to recognize the three-dimensional real environment from two-dimensional images. Computers can not only perceive the structure and motion information of three-dimensional objects, including its shape, position, posture, motion, etc., but also recognize and understand them. . In fact, most of Marr's computational theoretical framework and other computational theoretical frameworks formed in the 1980s involve using geometric methods to calculate the shape, position, attitude, and motion of three-dimensional objects in the environment. Existing camera calibration methods can be divided into traditional camera calibration methods and camera self-calibration methods. The traditional ...

AI Technical Summary

Problems solved by technology

[0003] The shooting range of traditional cameras is very limited, and the effective information obtained cannot meet people's needs, so there is the birth of panoramic cameras, among which panoramic cameras can be divided into central and non-central

Images