Consider a digital camera that faces an LCD and that is perfectly focused on the LCD. The LCD shows an image of $n \times n$ pixels. What resolution does the camera have to have to perfectly capture the image displayed on the LCD?
(Let's keep this simple and assume that camera pixels and LCD pixels are of the same size, pixels are "atomic" objects, no perspective distortion, etc.)
I'm asking because I have found the following statement in a paper:
"To satisfy the Nyquist criteria for image resolution, each pixel of the image shown on the LCD should be sampled by 2 or more pixels in the camera."
I don't understand this. Intuitively, I would say that if camera and LCD were perfectly aligned, then a camera with $n \times n$ pixels would suffice. My (poor) understanding of theory also suggests the statement is wrong, because every row or column of the image on the LCD can have a maximum frequency component of $n/2$. The minimum sampling rate would be $n$, not $2n$.
What am I missing?