研究目的
To develop a linear calibration method for paracatadioptric cameras using sphere images and common self-polar triangles to determine intrinsic parameters.
研究成果
The proposed method is feasible and effective for calibrating paracatadioptric cameras using sphere images and common self-polar triangles, providing stable results with less sensitivity to noise. It can be extended to other central catadioptric cameras with adjustments for the mirror parameter.
研究不足
The method is specifically for paracatadioptric cameras (ξ=1) and may not directly apply to other central catadioptric cameras without modification for the mirror parameter ξ. It relies on accurate extraction of sphere images and is sensitive to noise, though it shows improved stability compared to other methods.
1：Experimental Design and Method Selection:
The method involves projecting a sphere onto the unit viewing sphere to form antipodal circles, then onto the image plane to form conics. Common self-polar triangles between these conics are used to find the vanishing line and imaged circular points, enabling linear calibration of intrinsic parameters.
2：Sample Selection and Data Sources:
Simulated data with Gaussian noise and real images of a table tennis ball (sphere) taken from three different positions.
3：List of Experimental Equipment and Materials:
A paracatadioptric camera with a paraboloidal mirror, a table tennis ball as a calibration object, and a checkerboard for validation.
4：Experimental Procedures and Operational Workflow:
Extract edge points using Canny edge detection, fit conics using least-squares method, compute common self-polar triangles, determine vanishing line and imaged circular points, and solve for intrinsic parameters using constraints from three images.
5：Data Analysis Methods:
Comparison with existing methods (Ying and Duan) using absolute errors for intrinsic parameters in simulation and real experiments; validation through image rectification and Hough transformation for vanishing lines.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
