Sparse dictionary-based camera calibration method

Abstract

The invention discloses a sparse dictionary-based camera calibration method. The method comprises the following steps of: 1, shooting a plurality of images of a calibration plate at different heightsby using a camera, grouping data corresponding to a plurality of internal reference matrixes K into a data set Y; 2, randomly extracting k samples from the data set Y; 3, fixing a sparse dictionary and solving a sparse code; 4, updating the sparse dictionary D and a code matrix X column by column; 5, shooting an image of the calibration plate by using the camera, solving a sparse coefficient x, and solving an internal reference matrix K and an external reference matrix of the camera by utilizing the sparse coefficient x and the sparse dictionary D obtained in the step 4; and 6, solving a distortion coefficient, and carrying out maximum likelihood estimation to obtain optimum solutions of the distortion coefficient and the internal reference matrix K. According to the method, when each camera is used for the first time, a plurality of images of the calibration plate at different heights are shot by using the camera, and the sparse dictionary is constructed, so that calibration can be completed through shooting one image of the calibration plate during the using at each time in the future.

Description

technical field [0001] The invention relates to a camera calibration method, in particular to a camera calibration method based on a sparse dictionary. Background technique [0002] The early camera calibration technology originated from lens correction in photogrammetry in the nineteenth century, and was mainly used to solve the problem of precise matching between viewpoint coordinates and corresponding image point coordinates. The camera model is composed of parameters that define the matching relationship, and these parameters are usually divided into external parameters and internal parameters. The two-step calibration method proposed by Tsai in 1987 is considered to be an important theoretical work for camera calibration and has been widely used in the early camera calibration process. Based on the hypothetical model of collinearity between the radially distorted point and the real point, the method adopts a two-step strategy to solve the internal and external paramete...

AI Technical Summary

Problems solved by technology

In the process of taking multiple photos of the calibration board, not only is the operation cumbersome and the calibration efficiency is low, but also the angles of the calibration board may be placed repeatedly or have little difference due to human operation, which will lead to repeated calculation of useless data and make the calibration results unreliable. reduced accuracy

Images