# Lecture 21: Relative Orientation, Binocular Stereo, Structure, Quadrics, Calibration, Reprojection | Summary and Q&A

1.8K views
June 8, 2022
by
MIT OpenCourseWare
Lecture 21: Relative Orientation, Binocular Stereo, Structure, Quadrics, Calibration, Reprojection

## TL;DR

This content discusses the concept of relative orientation in photogrammetry and the process of camera calibration using a planar calibration target.

## Install to Summarize YouTube Videos and Get Transcripts

### Q: What are quadric surfaces and how are they classified in photogrammetry?

Quadric surfaces are a type of geometric shape used in photogrammetry. They are classified based on the number of negative signs in their equation, with positive signs representing ellipsoids, negative signs representing hyperboloids, and complex numbers representing imaginary ellipsoids.

### Q: How is camera calibration done using a planar calibration target?

Camera calibration involves finding the intrinsic parameters (such as the principal point and focal length) and extrinsic parameters (such as rotation and translation) of a camera. In the case of a planar calibration target, perspective projection equations are used along with known correspondences between 3D points on the target and their 2D image positions. Nonlinear optimization methods such as Levenberg-Marquardt are often used to minimize the error between the predicted and measured image positions.

### Q: What is the advantage of using a planar calibration target?

Planar calibration targets are easy to manufacture, store, and have high accuracy. They allow for precise measurement of 2D points, making calibration more straightforward compared to three-dimensional targets. However, depth variations are required to fully determine parameters such as focal length and translation.

### Q: How is noise sensitivity addressed in camera calibration?

Monte Carlo methods can be used to analyze noise sensitivity in camera calibration. By adding known statistical properties of noise to measured image positions and performing the calibration computations multiple times, the statistical properties of the resulting parameters can be examined. This allows for the assessment of the noise gain or sensitivity to measurement errors in the calibration process.

## Summary & Key Takeaways

• The content explains the concept of relative orientation and its application in photogrammetry, specifically in binocular stereo and motion vision structure.

• It discusses the classification of quadric surfaces and the circumstances under which the relative orientation cannot be determined.

• The content also introduces the process of camera calibration, including interior orientation and exterior orientation using a calibration target.