Non-Overlapping Multi-Camera Calibration Dataset using Icosahedron and Cube Calibration Object

Our dataset comprises images captured using two 3D calibration objects: an icosahedron and a cube. Six 20MP monochrome cameras were arranged in a semicircular configuration with minimal overlapping fields of view. A consistent robotic motion pattern was employed to acquire sequential images for both calibration objects.

This work is conducted at the Institute for Factory Automation and Production Systems(FAPS), University of Erlangen-Nuremberg, Germany.