| Summary: | Much like how buildings form the physical infrastructure of a city, 3D models of buildings and their interiors form the digital infrastructure of a smart city. Such models can be used for urban planning, facility management or even in search and rescue operations.
Google’s Project Tango, a tablet equipped with an RGB-IR camera, uses RGB-D technology to perform indoor mapping by capturing depth and colour data of each pixel in the scene. This study aims to evaluate the tablet’s ability to reconstruct indoor environments and to integrate the tablet with a Pioneer P3DX to form an indoor mapping system. Indoor mapping of three locations were done via three approaches: hand-held, mounted on a remote-controlled car and mounted on the P3DX robot.
In the hand-held approach, the tablet’s capabilities and limitations were evaluated. The resolution and accuracy of the mesh was of sufficient quality to give an overall representation of the environment but lacked precision on the smaller scale. Various situations which inhibited the tablet’s ability to perform mapping were also discussed.
The remote-controlled car approach served to highlight some of the problems and design considerations when mapping without a human operator. A field of view and depth study was also conducted to further understand the capabilities of the sensors.
The third approach involved the design and fabrication of a pan-tilt apparatus used to interface between the tablet and the P3DX robot. The resulting map generated by this mapping system was comparable to the one generated in the hand-held approach, demonstrating the suitability of the apparatus.
Overall, this project provides insights on the use of Project Tango and seeks to provide a platform for future work towards a fully autonomous mapping system.
|
|---|