Geospatial Cloud Computing - Perspectives for Startups and Open Source Projects The Cases of OpenWebGlobe and 3D Vision Mobile Mapping for Urban Infrastructure Management Prof. Dr. Stephan Nebiker FHNW University of Applied Sciences and Arts Northwestern Switzerland Institute of Geomatics Engineering, Muttenz, Switzerland Geospatial World Forum, Geneva, 9 th May 2014 2014 FHNW / IVGI
Motivation New geospatial technologies and solutions often create / produce / process / involve Big Data require massive storage capacities, processing power and in case of success high network bandwidths potentially massive financial investments Cloud Computing as a solution? 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 2
Key Factors for new Geospatial Startups Great new ideas mixed with enthusiasm & enterpreneurship Meet the expectations short time to market with minimal financial investments Don't disappoint also in the event of success! be prepared for rapid and massive scaling of storage, processing and exploitation capacities The answer Cloud Computing 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 3
Outline Introduction / Motivation OpenWebGlobe An Open Source Virtual Globe Technology Background and Vision Characteristics Solution infra3d 3D Vision Mobile Mapping for Urban Infrastructure Management Conclusion and Outlook 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 4
Virtual Globes & Startups earlier lessons learned 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 5
Virtual Globes & Startups earlier lessons learned (Updated) Vision Provide an interactive Virtual Globe technology supporting up-to-date, open and customer data with high-levels of detail on a global scale... without a Google budget! 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 6
3rd generation Virtual Globe without plugins for modern webbrowsers, using WebGL Open Source Project started in April 2011 (based on C++ version started in 2005) JavaScript Library for rapid development of web-based 3D geoinformation applications Full integration into customer web sites Runs on Windows, MacOS X, Linux Broad support for geospatial data types: imagery, maps, elevation, 3d objects, point clouds 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 7
Demo http://world.openwebglobe.org 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 8
Other Data Sources Architecture Solution #1: Preprocessing all tiles in the cloud WMS, WMTS, TMS,... Postgres / PostGIS 3DPS (in Future) download visible data fragments (tiles) Generate tiles (for Streaming) (Web-) Viewer HTML5 JavaScript WebGL for 3D Upload Raw data to Cloud (or web server) Raw: Image, Elevation, 3D- Models, Point Clouds,... 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 9
Architecture Solution #2: Partial Preprocessing Tiles & On-The-Fly Tile Generation and Caching Hybrid Cloud Solution 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 10
HPC: 120 TB Data (900MB/s), > 50 CPU Cores in combination with the Cloud 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 11
Showcase: Streaming 3D-Geometry Tiles (from OSM data & Buildings of Interest) Bachelor Thesis / Master Thesis @FHNW IVGI Image courtesy of BTh Hürbi/Daetwyler, MTh Lucas Oertli, 2013 MapData OpenStreetMap contributors 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 12
On-the-fly generation and streaming of 3D city models 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 13
Outline Introduction / Motivation OpenWebGlobe An Open Source Virtual Globe Technology infra3d 3D Vision Mobile Mapping for Urban Infrastructure Management Vision Characteristics Solution Conclusion and Outlook 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 14
The vast majority of Mobile Mapping relies on Laserscanning So why Imagery and why 3D Image-based Mobile Mapping? 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 15
An estimated 90% of Mobile Mapping relies on Laserscanning So why Imagery and why 3D Image-based Mobile Mapping? A picture is worth a thousand words! and Humans see and think in images not in point clouds! 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 16
Vision The 3D Geoinformation Loop based on Intelligent 3D Image Spaces Research Projects SmartMobileMapping (*) infravis (*) (*) supported by the Commission for Technology and Innovation CTI 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 17
3D Vision Mobile Mapping Acquisition Systems Stereo Cameras (oblique) Laser Scanner Mobile Platform GNSS RX Motor Vehicle, Aircraft, UAV, Train, Boat INS Multiple Stereo Systems high-performance Stereo industry cameras Cameras up to 11 MP, 5 to 30 fps up to 5 stereo systems (i.e. 10 cameras) Onboard data storage highly parallel 1-2 TB per hour 2014 FHNW / IVGI Multi-Stereo Mobile Mapping System by inovitas AG (www.invovitas.ch Geospatial World Forum 2014, Geneva 18
3D Vision Mobile Mapping Components and Workflow 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 19
3D Point Coordinate Determination Accuracy (MSc Project Cavegn, 2010; Bachelor Thesis Arnold, 2011) Empirical Standard Deviations of Coordinate Differences stereovision mobile mapping vs. tachymetry (reference) approx. 50 check points Position Height Absolute 3-4 cm 2 cm 3D 4 cm Relative 3D Distance < 1 cm With ground control points & integrated sensor orientation 1-2 cm absolute 3D accuracy (even in tunnels) 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 20
infra3d ASTRA / FEDRO Swiss Federal Roads Office 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 21
infra3dcity Lucerne, Switzerland 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 22
infra3drail bls (Brig Lötschberg Simplon Railway Lines) 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 23
2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 24
Conclusions Geospatial Cloud Computing Interesting perspectives for Startups and Open Source projects OpenWebGlobe (www.openwebglobe.org) Open Source technology / MIT license join the community! flexible support for private, public and hybrid cloud solutions 3D Vision Mobile Mapping for Urban Infrastructure Management streetview for professionals (3D, accuracy, integration) entirely cloud based apart from the real-world data capturing process infra3d services commercially available from inovitas AG (www.inovitas.ch) we are just at the beginning! 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 25
Thank you for the attention! Questions? Contacts Stephan Nebiker stephan.nebiker@fhnw.ch @snebiker OpenWebGlobe www.openwebglobe.org @openwebglobe infra3d / inovitas AG www.inovitas.ch 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 26
Important Tile Types 2D Image Tile contents: 256x256 Pixels (in most cases) (May also contain rasterized vectors etc.) 2D Vector Tile contents: 2D Geometry (Number of Elements limited for LOD) 2D Elevation Tile (-> for creating 3D Tiles) contents: Elevation values (e.g. numbers) contents 17x17 values (or similar). 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 27
3D Point Cloud Extraction using Dense Stereo- or Multi-Image Matching 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 28
Interactive 3D Exploitation 3D Imagery Repository Precise 3D Measurements in 3D Image Space Measurements based on 3D monoplotting Stereo Imagery Monoscopic Exploitation Depth Maps navigation data stream metadata Interactive measurements in 3D videos Georeferenced 3D Imagery / Videos Stereoscopic Exploitation 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 29
Conclusion Strengths of 3D Vision-based Mobile Mapping very dense stereo image sequences => redundancy, accuracy easy-to-use 3d environment / ideal for non-geospatial experts automatic 3d point measurements very powerful integrated georeferencing possibilities strong automatic object extraction capabilities possibility to use LiDAR as secondary sensor system 3D Vision Mobile Mapping & Cloud-based 3D Exploitation powerful, accurate and operational solution ideally supports the vision of "The Geoinformation Loop" and we are only at the beginning 2014 FHNW / IVGI Geospatial World Forum 2014, Geneva 30