The software used to get stereo rectificaiton parameters

[Mechazo11]

Hi @DapengFeng

In alpha.yaml, bob.yaml and carol.yaml files, the stereo rectification data are provided in the following format

#--------------------------------------------------------------------------------------------
# Stereo Rectification. Only if you need to pre-rectify the images.
#--------------------------------------------------------------------------------------------
LEFT.height: 1024
LEFT.width: 1224
LEFT.D: !!opencv-matrix
   rows: 1
   cols: 5
   dt: d
   data: [-0.052628389768906, 0.100832689624109, 0.000395112517197, 0.000362060505238, 0.0]
LEFT.K: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [1172.771733138536,    0.0,            617.941545715258, 
             0.0,            1172.364005984343, 507.672727295385, 
             0.0,               0.0,              1.0]
LEFT.R:  !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [1.0, 0.0, 0.0,
          0.0, 1.0, 0.0,
          0.0, 0.0, 1.0]
LEFT.Rf:  !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [1.0, 0.0, 0.0, 
          0.0, 1.0, 0.0, 
          0.0, 0.0, 1.0]
LEFT.P:  !!opencv-matrix
   rows: 3
   cols: 4
   dt: d
   data: [1172.771733138536,    0.0,            617.941545715258, 0.0,  
             0.0,            1172.364005984343, 507.672727295385, 0.0,  
             0.0,               0.0,              1.0,            0.0]

RIGHT.height: 1024
RIGHT.width: 1224
RIGHT.D: !!opencv-matrix
   rows: 1
   cols: 5
   dt: d
   data: [-0.052817970234810, 0.100073648217903, 0.000987319404990, 0.000078198322045, 0.0]
RIGHT.K: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [1172.549603999100,    0.0,            617.477465353483, 
             0.0,            1173.241586185178, 516.335518983269, 
             0.0,               0.0,              1.0]
RIGHT.R:  !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [0.999994846143267,   0.003159631643868,  -0.000569574209205,
         -0.003162491376236,   0.999982035625659,  -0.005091863534975,
          0.000553475564009,   0.005093638565765,   0.999986874169337]
RIGHT.P:  !!opencv-matrix
   rows: 3
   cols: 4
   dt: d
   data: [1172.549603999100,    0.0,              617.477465353483, -361.1747016345565 , 
             0.0,              1173.241586185178, 516.335518983269,   -5.2678945955122, 
             0.0,                 0.0,              1.0,              -1.5931129591554]

Which software / procedure was used to compute this data? Is it possible to get T_c1_c2 where c1 is the left camera frame and c2 is the right camera from directly from the provided data?

[DapengFeng]

The data was computed using MATLAB's Computer Vision Toolbox.

[Mechazo11]

Thank you very much @DapengFeng,.

Were the calibration images used with the MATLAB toolbox saved? I am trying to run S3E with ORB-SLAM3 and one of the key parameters required is the relative transformation that brings right camera to left's coordinate frame. If the calibration image data are made available, the Kaliber software may be used to retrieve $T_{c_1c_2}$. Given the baseline is known, this transformation would immediately make it possible to run ORB-SLAM3 for S3E datasets.