December 2007


Ontario research contributes to Virtual Earth 3D mapping

Interview with Dr. Vincent Tao

Profile:
Virtual Earth, Microsoft's live map service, is gaining a lot of attention, partly because of its innovative 3D, interactive capabilities, including photo-realistic 3D city models, leading-edge photogrammetry and computer vision technologies. The man leading the research and development is no stranger to ORION and Canada's research and education community. Dr. Vincent Tao, head of York University's GeoICT Lab, and founder and President of GeoTango International Corp - acquired by Microsoft in 2005 - has pioneered many of the technologies that are driving the next-generation applications such as Google Earth, Virtual Earth and real-time "sensor grids". He has led the development of several innovative technologies and products in the areas of 3D Earth Web visualization and exploration and 3D modeling from imagery. The former Canada Research Chair and Professor of Geospatial Information Technology at York University, he has authored over 200 technical papers and contributed to several books. He received his PhD from the University of Calgary in 1997.


Dr. Vincent Tao, Director of Virtual Earth and Local Search Business Unit, Microsoft; Canada Research Chair and Professor, Geospatial Information Technology, York University.

What is your role at Microsoft?

I am responsible for business and technology strategy for the Microsoft Virtual Earth program and have recently devoted more efforts to international markets.

Virtual Earth is a business unit under Microsoft On-line Services Group. Microsoft has invested a lot in this area and developed some truly innovative solutions in this space including Virtual Earth 2D, 3D and mobile. Most of these services are accessible from maps.live.com, a consumer based web portal.

In addition to day-to-day work, I am looking at how we can really scale the Virtual Earth technology to consumers and enterprise, government users, and how to make a solid business model to have the scale of economics. I also maintain strong interests in technology innovations and still keep some time commitment on technology research.



Virtual Earth 3D mapping image from above the CN Tower and downtown Toronto

To what extent do you maintain a collaborative research relationship with your lab, students and colleagues at York University?

Officially I am on leave from York University, but I still spend some of my time with my students and researchers at the York GeoICT Lab. I have graduated two PhD students and published one book last year.

I love being connected with York University. I will act as adjunct professor with York effective January, 2008. My Dean is very supportive and encouraged me to keep the Lab research. We have secured a full-time senior researcher who will manage the Lab and direct on-going research projects. With all the funding we have, the Lab is financially self-sustainable for the next couple of years. Though my time with the Lab is limited, I'd like to contribute my experience and help sustain the GeoICT Lab.



VE 3D mapping image of Niagara Falls

To what extent does Virtual Earth and Live Maps platforms reflect the application of research from GeoTango and the innovations from the York University lab?

Microsoft developed the Virtual Earth system internally. I believe that my knowledge and research experience in this area contributed to the development. As the first geomatics company acquired by Microsoft, our technology, and more importantly the expertise behind the technology have been recognized. In the fast paced knowledge industry, many acquisitions are centered on people not just technology.


The "GeoWeb" is becoming an increasingly vital part of the web and the development of next-generation and web 2-0 applications and services. To what extent has Ontario and Canada been pioneers in this field?

I started my research on '3D Digital Earth' in 1998. I believe we were one of the earliest research teams to focus on this area. I believe that 'persistency counts'. I have been pursuing this research for some seven years.

My passion is to integrate 'geospatial mapping technology' with the mainstream and to bring geoweb mapping to massive numbers of users. It is interesting, and to my surprise too, that I now have an opportunity to work with the largest IT company in the world.

There is no doubt that Canada is very advanced in geomatics research and technology innovations. Ontario is actually the place where the first GIS (Geographic Information System) was born. Canada contributed several breakthrough innovations in this space, for example, large geospatial database structures, 3D Web exploration, navigation and positioning, spatial modeling, and geospatial interoperability etc. This is widely recognized by our industry.



VE 3D mapping image of Parliament Hill, Ottawa

Virtual Earth is being extended beyond mapping capability to include "online infotainment" and social networking. How will that work? What can users expect to find when researching a particular locale (such as vacation destination)?

Conventional mapping services generally offer capabilities such as finding the location, providing driving directions, etc. One of the key concepts in Virtual Earth is to extend our mapping to infotainment, social media and local search services. This is a very powerful concept for users. What users are interested in is not about the map itself, but in information contained on the map (location).

The information must be visually entertaining. It can better help users find a vacation destination, a house for rent or even conduct a virtual tour of a place. Infotainment scenarios require a lot of investment in images and 3D browsing experiences, in addition to rich local content information (restaurants, landmarks, local business, tour info etc).

The "map" itself can be considered as a type of social media, where it carries a lot of location-based information. Virtual Earth has pioneered some innovative features in this area. We have developed a feature called 'Collection' where users can annotate things on a map and collect them for their own or public uses. For example, you can share your hiking experience by offering your hike trail, photos or comments or any web links. Other hiking fans will likely provide comments or add their own trails. From this example, one can see that the map becomes virtually a 'blog' space.

We believe that most knowledge about 'local' (nice restaurants and good dentists etc) is not available through the web crawling. Thus, social media becomes an effective vehicle to collecting local information in people's minds.


In what way does Virtual Earth differ from Google Earth? Is it fair to describe Virtual Earth and Google Earth as rivals?

Like in sports, having competitive players will really help the game itself. With Google, Yahoo and other players in this space, I am delighted that the technology innovations have been largely accelerated in the last 18 months. Compared to Google Earth and Google Maps, two separate products, Virtual Earth offers an integrated and browser-based environment for both 2D and 3D user experiences.

You can explore location or local information either in 2D mode or 3D without a need to switch between products.

In terms of content, Virtual Earth has pioneered photo-realistic 3D city models in our system with our leading-edge photogrammetry and computer vision technologies. You can tell the difference in exploring some major Canadian cities in 3D, like Toronto, Montreal or Niagara Falls. We are very advanced in bird's-eye imaging. Bird's-eye images are oblique and of extremely high resolution (about 15 cm). Currently, we are extending our bird's-eye coverage to Canada and Europe.

There are many differences in features. Virtual Earth has developed several innovations including real-time traffic integration; hill-shading terrain map visualization; user generated content collections; on-click routing and driving directions; and also street-side images and consumer photo 3D browsing capabilities, etc.

Another key difference is that Virtual Earth has a set of rich APIs (Application Programming Interface) that allows software developers and businesses to develop their own software systems, applications or on-line services. We would like to see thousands, if not millions of Virtual Earths live. This is very different than Google Earth's current offering (i.e. one Earth approach). Our goal is to empower our developers and allow them to control their applications.


What is the potential of Virtual Earth for researchers or educators?

The increasing availability of local information in addition to geographic data such as maps or images offers researchers and educators a great source for research and studies. Also the cost of using the mapping software platform and the local mapping data is dramatically reduced - actually free under a certain condition for Virtual Earth licensing - and even students can use them for their course projects. On the other hand, there are lots of research questions for us to address. Among many of them, intelligent and automated geospatial data updating, local information mining and aggregation, social mapping, virtual tourism and shopping, location commerce etc. are some of my interests.


What is on the horizon for Virtual Earth and interactive mapping services and applications? To what extent, for instance, do you anticipate the deployment of a world-wide, integrated "sensor grid"?

Geospatial information will become indistinguishable as it is weaved into the fabric of information and communication technology (ICT). I used to use the term, GeoICT, but I think that the 'geo' part will disappear eventually. The word 'geospatial' was created to differentiate our industry from others. As 'geo' becomes a ubiquitous component in the ICT, the geospatial industry will appreciate the unprecedented benefits from the integration, not differentiation.

From an information viewpoint, 'what, where, and when' (WWW) are the three fundamental elements for information organization and indexing. The keywords-based database, management and search technologies, originally designed for addressing 'what' problems, will be significantly enhanced or even re-architected to accommodate the queries that have 'where' and 'when' dimensions.

The information and communication systems, from software, to platforms and to devices, such as databases, operating systems, Web, phones, TVs, smart glasses etc, will become more location-intelligent so as to handle the location-based information intuitively.

The demand for more current or even more real-time information will be increasing. This will lead to the proliferation of sensor grid technology.


Thank you.

Learn more at http://maps.live.com and www.geoict.net.


Back to Headlines