Lidar and Multi-Camera Calibration and Fusion
Kalibrace a fúze lidaru a vícekamerového senzoru
Authors
Supervisors
Reviewers
Editors
Other contributors
Journal Title
Journal ISSN
Volume Title
Publisher
České vysoké učení technické v Praze
Czech Technical University in Prague
Czech Technical University in Prague
Date
Files
Abstract
Tato práce se zabývá vývojem metody pro systematickou fúzi mračna bodů naměřeného LiDARem s barevnou informací ze sady kamer. Pro zvýšení robustnosti vyvinutého algoritmu bylo implementováno několik filtračních metod. Navržené řešení je obecně použitelné pro všechny robotické systémy. Hlavní motivací je ale použití těchto senzorů na palubě bezpilotní helikoptéry, kde budou data fúzována online během letu inspekční mise. Jako jedna z hlavních výzev fúze dat z rozdílných senzorů byla specifikována, navržena a implementována také kalibrace vnějších parametrů LiDARu a vícekamerového senzoru. Pro nalezení přesné transformace mezi senzory je v této práci použit šachovnicový vzor pro kalibraci. Vyvinuté metody jsou analyzovány a jejich chování je testováno na datech ze simulace i reálného světa.
This thesis deals with developing a method for a systematic fusion of the point cloud data measured by LiDAR with colour information from a set of cameras. Several filtration methods have been implemented to increase the robustness of the developed algorithm. The proposed solution is generally applicable for all robotic solutions. However, the main motivation is to equip all these sensors onboard an unmanned aerial vehicle and fuse the colour and the spatial information online mid-flight during an inspection mission. A process of calibration of the LiDAR and multi-camera sensors is also designed and implemented as one of the main challenges of the fusion process. With attention to finding the precise transformation, a calibration checkerboard pattern is used in this thesis. The developed methods are analysed and their performance is evaluated on data from simulation and real-world tests.
This thesis deals with developing a method for a systematic fusion of the point cloud data measured by LiDAR with colour information from a set of cameras. Several filtration methods have been implemented to increase the robustness of the developed algorithm. The proposed solution is generally applicable for all robotic solutions. However, the main motivation is to equip all these sensors onboard an unmanned aerial vehicle and fuse the colour and the spatial information online mid-flight during an inspection mission. A process of calibration of the LiDAR and multi-camera sensors is also designed and implemented as one of the main challenges of the fusion process. With attention to finding the precise transformation, a calibration checkerboard pattern is used in this thesis. The developed methods are analysed and their performance is evaluated on data from simulation and real-world tests.