Data fusion of lidar and RGB cameras
Tutor / directorRoyo Royo, Santiago
Document typeMaster thesis
Rights accessRestricted access - confidentiality agreement (embargoed until 2028-11-11T15:28:57Z)
All rights reserved. This work is protected by the corresponding intellectual and industrial property rights. Without prejudice to any existing legal exemptions, reproduction, distribution, public communication or transformation of this work are prohibited without permission of the copyright holder
Light Detection and Ranging - LiDAR or LIDAR - has traditionally been used for making high-resolution maps in a huge variety of geological applications due to its 2D and 3D point cloud mappings. Additionally, it has recently been found to be a crucial sensor in other applications such as autonomous driving. In several applications, LiDAR images are used in combination with conventional RGB images using a data fusion approach known as image registration. Thus, the main aim of this thesis is to combine both LiDAR point clouds and RGB images to provide color information overlapped to the 3D map. Hence, this work presents an image processing technique for registering a conventional color image with a point cloud. It was experimentally carried out with two different types of point clouds generated by a TOF and a LiDAR camera, the last one providing a new type of high-density point cloud. Thus, robustness of the technique for any point cloud and feasibility for high-density ones are proven.
Lidar imaging is a powerful measurement technique where a laser pulse is shone onto an object and the beam reflected back is recovered at some solid-state detector. The time elapsed is counted so an automated measurement of the distance to the target is obtained, without any further calculation. The concept is also referred to as ladar or time of-flight imaging. Most popular recent applications involve landing aids, object recognition, self-guided vehicles and safety and security applications in transport. In most applications, the system includes more than one imaging