Child pages
  • z_*OBSOLETE* Camera Calibration
Skip to end of metadata
Go to start of metadata


CRITICAL UPDATE: Use new XMALab software for marker-based XROMM, including Undistortion, Calibration, Marker Tracking, and Rigid Bodies.
XMALab replaces MATLAB XrayProject. See Bitbucket XMALab wiki for XMALab User Manual (not this wiki). 


  1. Open calibration tool
  2. Add camera
  3. Click the button next to Images
  4. Select all images and click Open
  5. Click the checkbox "Use Undistortion LUT"
  6. Click the "..." button and select the "_UNDSTFORM.csv" file for the camera
  7. Click "Add Camera"
  8. Repeat steps 1-7 for the 2nd camera
  9. Load a calibration target
  10. Select the point file for your cube (*.csv framespec)
  11. Mark the checkbox Reference points and select the reference file (*.ref) for your cube
  12. Load the points
  13. Now mark the reference points in 1 image of the first camera by clicking onto the image. The order and the reference points are shown in the sidebar.
    Make sure that you always select the marked point. (You can also select a point in the table to set it again in case you have to place them better.
  14. After you select the 4 points, it should look like this.
  15. Now you can compute the camera and the position of the camera by clicking "Init correspondences ---> by computing camera and pose (selected frame)"
  16. Now the points should be selected and the camera parameters in the sidebar set. In case it did not work, try running the same action again (step 15) or replace the reference points (step 13).
  17. Now repeat steps 13-16 for the 2nd camera. You can switch the camera by clicking on the tabs.
  18. Now you can compute the position for another image of the sequence. You can scroll through the images using the slider.
  19. First Select again 4 reference points as done in step 13 (3 or 2 can also be sufficient)
  20. After you have the points selected, you can run "Init correspondences ---> by computing pose with calibrated camera data (selected frame)".
  21. Now you can set the 2nd camera automatically by clicking “Init correspondences ---> by computing pose with calibrated cameras and pose data (all frames/cameras)”.
  22. Then you repeat step 19 to 21 until all images have correspondences established.
  23. Now you can go through all frames of a camera and check/uncheck points by either double clicking on them in the image or by changing the checkbox in the sidebar. All correctly detected points should be marked/green while points which are not accurately detected should be unmarked/red.Note: Blue points are back projected points of the calibration object but were not detected in the image.
  24. After all points are selected in one camera, you can update the camera parameters by clicking “Calibrate camera based on Inlier ---> Camera and pose (all frames)”.
  25. Then repeat step 23 and 24 for the other camera(s)
  26. Save the output by clicking one of the save options.
    1. DLT - Saves parameters computed with a DLT method.
    2. Model based assumes a general camera model. As of June 2014, this is the recommended option.
    3. Model based single focus assumes a general camera model with a focal length which does not differ in x and y. This is the maya camera model.


Unknown macro: {center}                 Table of Contents                

Unknown macro: {center}                      Workflow Diagrams