Validation Methods for Geometric Camera Calibration

Name: Validation Methods for Geometric Camera Calibration
Uploaded: 2020-03-23T00:00:00+00:00
Duration: 18 min 5 s
Description: Paul Romanczyk, Senior Imaging Scientist at Imatest, presents his Electronic Imaging 2020 paper. Camera-based advanced driver-assistance systems (ADAS) need the mapping from image coordinates to world coordinates, which geometric calibration provides. The talk describes tests for judging how good a...

Mar 23, 2020 Image Quality Factors Industry Events

Paul Romanczyk, Senior Imaging Scientist at Imatest, presents his Electronic Imaging 2020 paper. Camera-based advanced driver-assistance systems (ADAS) need the mapping from image coordinates to world coordinates, which geometric calibration provides. The talk describes tests for judging how good a calibration is and for comparing calibration models for an as-built camera, driven by automotive use cases:

Image coordinate system test: are different teams using the same image coordinates?
Reprojection test: forward-project targets through the model onto the image plane
Projection test: inverse-project points through the model out into the world
Triangulation test: how well a multi-camera system locates a point in 3D

Download the paper and slides in Validation Methods for Geometric Camera Calibration. Also on YouTube.

Download this video