In Chapter 7, the transformation matrix results obtained by different algorithms are different. What is the reason?

[zyqdragon]

In Chapter 7, the transformation matrix results obtained by solving three different algorithms of product geometry/icp/pnp are different. What is the reason?

1. R and t obtained by solving the epipolar geometry are as follows:
   R is
   [0.9985961798781875, -0.05169917220143666, 0.011526713598279;
   0.05139607508976057, 0.9983603445075082, 0.02520051547522394;
   -0.01281065954813587, -0.02457271064688466, 0.9996159607036126]
   t is
   [-0.8220841067933343;
   -0.03269742706405145;
   0.5684264241053513]

2. The transformation matrix obtained by solving the pnp algorithm is as follows:
   0.997866202583 -0.0516724161336 0.0399124436024 -0.127225965696
   0.050595891596 0.998339762772 0.0275276919261 -0.00750729765631
   -0.04126860183 -0.0254495477384 0.998823919929 0.0613858417711
   0 0 0 1

3. The transformation matrix obtained by solving the ICP algorithm is as follows:
   0.99724 0.0561704 -0.04856 0.141725
   -0.0559835 0.998418 0.00520242 -0.0555103
   0.0487754 -0.0024695 0.998807 -0.0311913
   0 0 0 1