Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let's assume we have a linear coordinate system on a plane. If we make a 2d perspective projection of that plane from 3d, we get a "skewed" coordinate system.

For example, one axis would look like this in the original plane:


but will look something like this in the projection:


I don't have any information at all about the focal length, the viewing angle, the distances involved, anything. The only information I have are the distances between the "markings" of the new coordinate system, and the knowledge that it was linear before the projection. this is why I can't find a way to use the well-known formulas, matrix notations, etc. of the perspective projection.

If I knew what form this function can take, I would be able to calculate its parameters from the points I have. For example, if it were quadratic, in the form of $ax^2 + bx + c$ I could calculate $a$, $b$ and $c$ by using the data points I have. I know it's not quadratic, so what can it be?

share|cite|improve this question
Where does this problem come from? According to your description, is it from 3D vision field? If yes, there are a mature mathematical system on 3D vision and of course perspective projection. – Shiyu Dec 1 '12 at 16:07
up vote 1 down vote accepted

A perspective projection of a plane is a special case of a projective transformation. So you can model it as a $3\times 3$ matrix operating on homogenous coordinates. Such a transformation is fully defined if you know the preimage and image coordinates for four points.

For points on a single line, like the ones in your illustration, three points are sufficient. The cross ratio is a particularly useful tool to reconstruct a scale on a projectively transformed line.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.