A Novel Camera Calibration Method Based on Multilevel-Edge-Fitting Ellipse-Shaped Analytical Model

In this paper, a novel camera calibration method based on a multilevel-edge-fitting ellipse-shaped analytical model (CCM-MEF-ESAM) is proposed. First, an ellipse-shaped analytical model (ESAM) is derived which considers distortion error of the spherical image. Second, multilevel-edge-fitting (MEF) is presented to select the edge points to fit the ellipse. Then the actual projection of the sphere is calculated by CCM-MEF-ESAM, which can be utilized to obtain the accurate external parameters of the camera. Finally, the robustness and accuracy of CCM-MEF-ESAM are validated by simulations and experiments. Within the environments of incomplete edges and complex illumination, CCM-MEF-ESAM can realize an improvement of calibration accuracy of 74% compared to the traditional method, which indicates that CCM-MEF-ESAM holds great potential in the field of vision navigation.

I. Introduction

Camera calibration is one of the most important steps in machine vision navigation, which directly determines the accuracy of a system [1]. Especially in a medical navigation system, an operation requires high precision in a complex operating room environment, which poses a great challenge to camera calibration [2].