I've created [X,Y,Z] co-ordinates for all cameras using the dot-product of: a camera's transformation matrix, and the 'centre' [x,y,z,1] matrix that are both located in an exported cameras.xml file (exported camera calibration).
Is this a correct approach to re-calculating each camera's co-ordinates in 3D space? -Relative to the model?