Ecole Centrale of Nantes
Computer Vision (2 hours, closed books)
Part I
1 Generalities
1. Explain the differences between identi cation and calibration.
2. Explain the general procedure to complete camera calibration (give a owchart and explain with
sufficient details).
3. Consider the openCV model for camera. What is the meaning of parameter k2 and k3? What
are their effect in image plane?
4. What represent fu and fv. Give details?
5. Explain what is a " perspective projection", an " orthographic projection", and a " paraper-
spective projection"? Give details and scheme.
2 Camera calibration
2.1 Linear method
Explain the principle of the method. Give details when using a pinhole model for camera.
2.2 Nonlinear method
Explain quickly the nonlinear method, then answer the following questions:
1. What is the meaning of reprojection error? Give a simple expression?
2. How many parameter you must estimate when using the OpenCv model for camera with a known
calibration board? List and describe all of them.
3. How many parameter you must estimate when using the OpenCV model for camera with an
unknown calibration board?
4. In that cases, how many points you need if you take 4 images? Give details and explanation.
3 Pose estimation
1. Explain what is a "Scaled orthographic projection"?
2. In which algorithm you can nd this projection?
3. What is the minimumn of points to estimate the pose of a known object? Explain why?