An Accurate 1D Camera Calibration Based on Weighted Similar-Invariant Linear Algorithm
In recent years, researchers around the world have been researching and improving the technique of 1D calibration of cameras. The previous work has been primarily focused on reducing the motion constraints of one-dimensional calibration objects, however the accuracy of the existing methods still needs to be improved when random noise is introduces. In order to improve the accuracy of the one-dimensional calibration of the camera, in this paper, we propose a new calibration method by combining a weighted similar invariant linear algorithm and an improved nonlinear optimization algorithm. Specifically, we use the weighted similar invariant linear algorithm to obtain the camera parameters as the initial calibration parameters, and then optimize the parameters by using improved nonlinear algorithm. Finally, in the case of introducing random noise, the results of computer simulations and laboratory experiments show that when the noise level reaches 2 pixels, the parameter error of this method is mostly reduced to 0.2% compared with other methods, which verifies the feasibility of our proposed method.
KeywordsCamera calibration Linear algorithm Nonlinear optimization 1D calibration objects
This work is financially supported in parts by the Fujian Provincial Department of Science and Technology of China (Grant No. 2019H0006 and 2018J01774), the National Natural Science Foundation of China (Grant No. 61601127), and the Foundation of Fujian Provincial Department of Industry and Information Technology of China (Grant No. 82318075).
- 3.Yoo, J.S., et al.: Improved LiDAR-camera calibration using marker detection based on 3D plane extraction 13(6), 2530–2544 (2018)Google Scholar
- 6.Zhang, Z.Y.: Flexible camera calibration by viewing a plane from unknown orientation. In: IEEE International Conference on Computer Vision (1999)Google Scholar
- 11.Akkad, N.E., Merras, M., et al.: Camera self-calibration with varying intrinsic parameters by an unknown three-dimensional scene. Int. J. Comput. Graphics. 30(5), 519–530 (2014)Google Scholar
- 13.Finsterle, S., Kowalsky, M.B.: A truncated Levenberg-Marquardt algorithm for the calibration of highly parameterized nonlinear models 37(5), 731–738 (2011)Google Scholar
- 17.Wang, L., Duan, F.: Zhang’s one-dimensional calibration revisited with the heteroscedastic error-in-variables model. In: 8th IEEE International Conference on Image Processing. IEEE (2011)Google Scholar