“…If not, the x c , y c , z c , f, [T ] determined from equations ( 8), ( 6) and (7) and the (ṽ i , wi ) calculated from equation ( 12) are inaccurate, and then the distortion model derived from equations ( 17) and ( 18) cannot be accurate. After the model parameters K 1 , K 2 , K 3 , P 1 and P 2 are obtained, one can use this improved lens distortion model to repeat the whole signal processing steps shown in equations ( 6)-( 8) and (10) to have more accurate estimations of x c , y c , z c , f, T 11 , T 12 and φ for each camera.…”