Building three-dimensional object models using a mobile platform and an RGBD camera

Damjan Tratnik (2013) Building three-dimensional object models using a mobile platform and an RGBD camera. EngD thesis.

Preview

PDF Download (15Mb)

Abstract

Nowadays we can observe 3D models in computer games, computer enriched movies, augmented reality and animations, as well as in the designing of new products, simulations, architecture, medicine etc. Due to the arrival of new low-cost depth cameras (e.g. Microsoft Kinect and Asus Xtion Pro), constructing 3D models of objects has become much more accessible and widespread. In this diploma thesis, we have prepared a short review of several distinct techniques of capturing shapes of objects and constructing their 3D models. Additionally, we implemented some of these techniques in a system that we designed. The system is based on mobile robot platform TurtleBot, which is composed of color-depth camera Microsoft Kinect and robot vacuum cleaner iRobot Roomba. Our software framework is based on Robot operating system (ROS) and Point Cloud Library (PCL), a software library for processing point clouds. Our system is able to automatically capture the 3D information of a target object and build its 3D model by using iterative closest point (ICP) algorithm. The 3D model obtained by the algorithm can be saved as a colored point cloud or as a polygon mesh.

Actions (login required)