Distortionsparametrar
Distortionsparametrar are numerical values that quantify deviations of an imaging system's ideal projection from reality, caused by lens and sensor imperfections or by transmission effects. In computer vision and photography, they particularly describe lens distortion, which alters the mapping from 3D world points to 2D image coordinates. The standard distortion model separates radial distortion, which projects points toward or away from the image center, and tangential distortion, which shifts points due to lens-sensor misalignment. The radial component is usually described by coefficients k1, k2, k3 (and sometimes higher-order terms), while the tangential component uses p1 and p2. In practice the distortion is described together with the intrinsic camera parameters (focal lengths, principal point, skew) in a calibration model, commonly the Brown–Conrady model or its variants. Some models also include a division model or additional higher-order terms for improved accuracy.
Estimation of distortionsparametrar is typically performed during camera calibration. A set of known patterns, such as
Limitations include model choice, sensitivity to calibration data, and non-uniqueness of solutions. Distortionsparametrar describe a simplified