![]() |
AliceVision
Photogrammetric Computer Vision Framework
|
Abstract class to represent the inverse operation of a distortion model. More...
#include <Undistortion.hpp>
Public Member Functions | |
| Undistortion (int width, int height) | |
| virtual EUNDISTORTION | getType () const =0 |
| virtual Undistortion * | clone () const =0 |
| bool | operator== (const Undistortion &other) const |
| void | setDesqueezed (bool isDesqueezed) |
| bool | isDesqueezed () const |
| bool | isLocked () const |
| void | setLocked (bool lock) |
| void | setOffset (const Vec2 &offset) |
| void | setSize (int width, int height) |
| void | setDiagonal (double diagonal) |
| void | setPixelAspectRatio (double pixelAspectRatio) |
| Vec2 | getOffset () const |
| Vec2 | getScaledOffset () const |
| Vec2 | getSize () const |
| Vec2 | getCenter () const |
| double | getDiagonal () const |
| double | getPixelAspectRatio () const |
| const std::vector< double > & | getParameters () const |
| void | setParameters (const std::vector< double > ¶ms) |
| std::size_t | getUndistortionParametersCount () const |
| Vec2 | undistort (const Vec2 &p) const |
| Eigen::Matrix< double, 2, Eigen::Dynamic > | getDerivativeUndistortWrtParameters (const Vec2 &p) |
| Eigen::Matrix< double, 2, 2 > | getDerivativeUndistortWrtOffset (const Vec2 &p) |
| Vec2 | inverse (const Vec2 &p) const |
| add distortion (return p' such that undisto(p') = p) | |
| virtual Vec2 | inverseNormalized (const Vec2 &p) const =0 |
| virtual Vec2 | undistortNormalized (const Vec2 &p) const =0 |
| virtual Eigen::Matrix< double, 2, Eigen::Dynamic > | getDerivativeUndistortNormalizedwrtParameters (const Vec2 &p) const =0 |
| virtual Eigen::Matrix< double, 2, 2 > | getDerivativeUndistortNormalizedwrtPoint (const Vec2 &p) const =0 |
Protected Attributes | |
| Vec2 | _size |
| Vec2 | _center |
| Vec2 | _offset |
| double | _diagonal |
| double | _pixelAspectRatio |
| bool | _isDesqueezed |
| std::vector< double > | _undistortionParams {} |
| bool | _isLocked |
Abstract class to represent the inverse operation of a distortion model.
Contrary to distortion models, the undistortion process is a pixel-to-pixel operation and thus it does not depend on the camera's focal length.
Its main purpose is to undistort images before they are used in other calculations so that we can virtually "remove" distortion parameters from the camera.
1.8.17