Skip to main content
A camera turns light from a 3D scene into a 2D array of pixel values. To recover geometry from those pixels, you need a model of that mapping, and no real camera matches the simplest model exactly. The sections in this chapter build the model in three steps.

Pinhole Camera Model

The pinhole projection, the intrinsic matrix, the lens distortion model, and coordinate conventions in ROS 2.

How Lenses Form Images

Snell’s law, the lensmaker formula, conjugate points, depth of field, concave lenses, and the Galilean telescope.

From World Points to Pixels

The full camera model with intrinsic and extrinsic parameters, backprojection, the horizon, and calibration from known points and from planar targets.
How the resulting images are represented and processed, from pixel arrays to filtering, is the subject of the Image Processing chapter.
Edit this page on GitHub or file an issue.