Lecture 4: Transforms (63)
wilrothman

Kind of a dumb thought but can this be done in polar coordinates instead of cartesian coordinates? Or would it violate convention so much to the point where it is generally outright rejected? Maybe we can interchangeably translate between a cartesian camera and a polar camera.

Furthermore, doesn't this need to incorporate all three types of rotation?

brandonlouie

I understand that the z-axis is negative looking down the camera is a result of the right hand rule. I think one possible way to generate a z-axis that is positive looking down the camera is to have the x-axis point the opposite direction. I wonder why this convention wasn't adopted instead? I supposed maybe it's a result of trying to maintain the 2D convention of having the positive x-axis be pointing right and the positive y-axis be pointing up if you were behind the camera (with the POV of the photographer). Although, if you were in front of the camera (with the POV of the photo subject), then the positive x-axis will now point to the left, which is against the convention. In this case, my suggested axes would work better (and I personally find them more intuitive). I guess in the end, someone decided that the POV of the photographer was more important and the convention on the slide stuck. It's interesting to think about how this came to be :)

You must be enrolled in the course to comment