Advances in Computer Vision

Image Formation

Week 1, Thursday

Pinhole Cameras

-Why don’t you see an image when you hold up a piece of paper? - Each point in space reflects in all directions, which get spread out across the paper - Conversely, each point on the paper gets light from multiple points in the scene.

• Pinhole camera light in such that each point on the image corresponds to a point from an object
• Get an image on the other side
• Tradeoff between brightness and sharpness

  Derivation

• Establish 3D camera system.
• center = pinhole cetner.
• z = distance from pinhole center
• x and y are up and side
• right handed in real-world space
• f = distance from pinhole to image plance

To do: better notes for lectures 1 (introduction), 2 (image formation), and 3 (fourier transform).