32 minutes ago · 5 min read1076 words · Tech · hide

The Nyquist criterion is frequently invoked as a limit on the detail that can be extracted from a digital image. That’s appropriate if we’re trying to reconstruct an arbitrary scene from its samples, but it can be overly restrictive when we already know something about the subject. Prior knowledge can allow us to estimate geometric properties at scales smaller than either the diffraction limit or the pixel spacing. Consider a simple target consisting of two horizontal, opaque black lines on a white background. The lines have known, equal widths, perfectly straight edges, and step-function transitions from black to white. They’re separated by a gap whose width we want to measure. We image them with a diffraction-limited lens and a digital sensor. For now, let’s arrange the lines parallel to the rows of pixels. The image formed by the lens is not a faithful reproduction of those step functions. Diffraction spreads the light, blurring the edges of the lines. As the lines approach each…

Commenting on Bubbles has been removed. Why? If you want to talk about this post, write to the author, use the comments on their blog if it has them, or reply with a "RE:" post on your own blog. Suggestions? feedback@bubbles.town