Camera Calibration Based On The Projection Property Of Circles

Camera calibration has attracted more and more attention,for it is important in burgeoning computer vision field.Because circles have good projective properties,they are studied in camera calibration.This thesis studies the calibration method of pinhole camera.On the one hand.Firstly,dual projection equation of the circle is explained by double contact relation geometrically.Secondly,we consider the three coplanar circles on the template in pairs,the generalized eigenvectors corresponding to the dual images of the two circles can be obtained,which is a line connecting the images of two centers,and the images of circle centers are obtained.Thirdly,the image of the circular points and vanishing line can be obtained from the image of circle centers and dual image of circular points is recovered.On the other hand.Firstly,a set of quadratic equations can be established by using the degenerate conic represented by the image of the circle center.Secondly,we consider the three coplanar circles on the template in pairs,the homogeneous equations constraint on the dual image of circular points is obtained algebraically,then the dual image of circular points is recovered.According to the property that the dual image of circular points encodes the homography matrix from world plane to image plane,the internal and external parameters of the camera are solved.At last,the proposed algorithm is evaluated on MATLAB R2017 b experimental platform,and the results show that the effectiveness of the raised algorithm.