Hello all hope everyone is doing fine.
I have a vision related problem where I want to calibrate my camera using a 3d object such as rubik's cube with known dimensions. I have a cube with 60mm length, width and height. I want to find the projection matrix "P". Which gives me the relation between 2D and 3D coordinates i.e. image taken from camera and the real word coordinates.
The formula used for this is: x=PX where x is 2D matrix of 3xn, P is projection matrix of 3x4 and X is 3D of 4xn.
one such example for x is:
Example for X is:
After finding that P matix of 3x4 something like this:
I use the inverse formula for finding the 3D coordinates:
inv(P)x = X
But inverse of P can't be be calculated so pseudo inverse of matrix is calculated using the pinv() command in matlab.
After calculating all these values I cant get back the right projection for 2D to 3D points using this matrix. is there something I am missing or I don't know yet to apply this? Also if you could refer me to a book or article where I could find its references.
I found these following refrences:
- Computer Vision:Algorithms and Applications by Richard Szeliski
- Multiple View Geometry in computer vision by richard Hartley and andrew Zisserman
