Expand description
This crate seamlessly plugs into cv-core
and provides pinhole camera models with and without distortion correction.
It can be used to convert image coordinates into real 3d direction vectors (called bearings) pointing towards where
the light came from that hit that pixel. It can also be used to convert backwards from the 3d back to the 2d
using the uncalibrate
method from the cv_core::CameraModel
trait.
Structs§
- This contains intrinsic camera parameters as per this Wikipedia page.
- This contains intrinsic camera parameters as per this Wikipedia page.
- This contains basic camera specifications that one could find on a manufacturer’s website. This only contains parameters that cannot be changed about a camera. The focal length is not included since that can typically be changed and images can also be magnified.
- This stores an essential matrix, which is satisfied by the following constraint:
- A point in normalized image coordinates. This keypoint has been corrected for distortion and normalized based on the camrea intrinsic matrix. Please note that the intrinsic matrix accounts for the natural focal length and any magnification to the image. Ultimately, the key points must be represented by their position on the camera sensor and normalized to the focal length of the camera.
- Allows pose solving to be parameterized if defaults are not sufficient.
Functions§
- Find the reprojection error in focal lengths of a feature match and a relative pose using the given triangulator.