Camera models

A camera model describes the relationship between 3D world and 2D pixel coordinates. Model parameters can be specified with supported files, or directly with the camera API.

Frame cameras

Orthority uses the term frame camera for area-scan cameras like those used in drone and aerial surveys. These cameras capture a 2D image in a single instance of time. The frame camera model is a physical model that splits the image formation process into interior and exterior components, each with its own set of parameters. In a typical drone or aerial survey, interior parameters are fixed for a camera, and exterior parameters vary for each image.

Interior parameters

The interior component describes the relationship between camera and pixel coordinates. It depends on the interior geometry and optical properties of the camera. Interior parameters are: image size, focal length, sensor size, principal point and distortion coefficients.

Exterior parameters

The exterior component describes the relationship between world and camera coordinates. It is an affine transform consisting of a translation and rotation. Orthority represents exterior parameters as a world coordinate (x, y, z) camera position, and (omega, phi, kappa) camera orientation, where the (omega, phi, kappa) angles rotate from camera to world coordinates.

RPC cameras

The RPC (rational polynomial coefficient) camera model represents the relationship between world and pixel coordinates as the ratio of two polynomials. It is a general (non-physical) model typically used for describing narrow field of view and push-broom satellite cameras.