Monocular camera calibration based on phased target and neural network
;Aiming at the problems of low calibration accuracy and large depth direction error of traditional calibration methods based on BP neural network,a monocular camera calibration method based on phase target and Radial basis function(RBF)neural network is proposed.The feature extraction method adopts a three-step phase shift method,and the multi frequency method is used to calculate the absolute phase.The absolute phase carried by the phase target feature points is converted into three-dimensional space,and the corresponding relationship between the feature point image coor-dinates and world coordinates is established.Finally,the RBF neural network is used to directly map the two-dimensional image coordinates to the three-dimensional space coordinates.The experimental results show that compared with the tra-ditional use of BP neural network for calibration and the calibration results of checkerboard and circular calibration tar-gets,the average calibration error of this method is 0.0980mm,At the same time,under a focal length of 1.4mm and a 220 ° field of view fisheye lens,high accuracy can still be maintained,proving the feasibility and effectiveness of the pro-posed method.
NETL
NSTL
万方数据
