Markerless AD-HOC calibration of a hyperspectral camera and a 3D laser scanner

Abstract

Integrating 3D data with hyperspectral images opens up novel approaches for several robotic tasks. To that end, we register hyperspectral panoramas to cylindrically projected laser scans. With our approach, the required calibration can be done on board a mobile robot without the need of external markers using Mutual Information. Qualitative results show the robustness of the presented approach, and an application example demonstrates possible future applications for hyperspectral point clouds.