Geography 2.0: Virtual Globes

AAG2007 Virtual Globes Logo

Thursday, August 24, 2006

Input an email, output to Google Earth

Originally coded as a Sci Foo party favor, follow this link (turnhole.com/rapid) to see my proof-of-concept implementation of a mobile Google Earth KML maker. I've been meaning to add error handling to the code, but have been busy on other things. Since it's too much fun to keep to myself, I thought I'd throw the idea and code into the wild, and see what the world does with it. Enjoy. ~Alan Glennon

Tuesday, August 22, 2006

Collaborative KML and Neven Vision

While coding a spreadsheet-to-KML parser for the online spreadsheet EditGrid, I noticed that they created one themselves. You'll need to register for a free account to access the tool (via My Workspace -> Add-ons -> Grid2Map). The EditGrid developers have also exposed the source code! Two weeks ago, Ogle Earth blogged about users calling individual EditGrid cell values from a Google Earth Networklink, but the latest tool is much quicker and more elegant. With the user designed tools, all the URL/cell queries really slow things down. Here's my code if you want to see how cells are called.

So, here's an experimental spreadsheet, and the associated KMZ. I didn't add a refresh rate, so either set one yourself or refresh manually by right-clicking over the file name in the Places menu. I suggest giving the KMZ a few seconds after the spreadsheet saves for the geocoder to work.

The big deal with these online spreadsheets is that they provide an inexpensive mechanism for multiple simultaneous user spatial data collaboration. Such spreadsheets are also a natural gateway to increased analytical functionality within virtual globes. Significant.



Neven Vision
Also, I read that Google acquired Neven Vision. Neven Vision writes computer vision software to automatically extract information from photos. The company had previously defined a key goal as becoming "the standard for machine vision on mobile phones." The company seems to be relatively well-established with about twenty thousand online references before the Google acquisition. Though the Neven Vision website has effectively been taken down, the Google cache reveals some information. For instance, Neven Vision has a product called iScout which allows a camera phone to "snap photos to initiate a search process and get relevant information or content sent" back to the phone. Some technical information on their technology to detect faces in low-resolution images can be found in this PowerPoint presentation (ppt).

Friday, August 18, 2006

Inside data

Concerning indoors spatial data, Mike Goodchild mentioned CityGML. References to interior models, which the University of Bonn initiative tags "LoD4" (level of detail 4), are not obvious on the main entry page, but are described in their presentation slides (pdf) and articles. Besides the CityGML site itself, Directions Magazine has an overview of the effort.

Thursday, August 17, 2006

KML mobile and Photosynth for indoors

As usual, the blog train rolls on: this blog posted about an entry on this blog and now you're reading it here. Anyway, the point is that Google Maps for Mobile now supports KML (link).

Changing topics
Concerning Photosynth, one interesting avenue may be the inside of buildings. As I ate lunch today, I thought about some work I did on 3D cave mapping in 1999. It involved me taking images and survey data to create 3D walkthroughs (side note: the old webpage about the project is still online at my former university). The thing that struck me is that years later, collection and organization of multiple data sources for the inside of real places has not seemed to change that much. Photosynth seems particularly appropriate. Eventually, we can start talking accuracy...

Wednesday, August 16, 2006

Microsoft Photosynth

All the standard sources have covered Microsoft Photosynth, but I wanted to mention a few additional links. For those who have not heard about it, Photosynth is a Microsoft Live Labs product that takes a collection of photographs, finds similarities, and arranges them in three-dimensional space.

As Stefan Geens at Ogle Earth points out, the technology "certainly looks like a prime candidate for inclusion in Microsoft's upcoming 3D virtual globe." Frank Taylor's Google Earth Blog has a long entry with similar comments. For an overview, the Microsoft Photosynth site has a four minute promotional video. However, to really get a feel for what Microsoft is doing, it is worth watching the 38-minute-long Channel 9 interview with Photosynth Architect Blaise Aguera y Arcas. Microsoft Research has a website on the technology with a number of compelling screenshots.

Image: Microsoft Corporation, 2006

Image: Microsoft Corporation, 2006


For an interactive preview of the concepts, the University of Washington also has a webpage on Photo Tourism technologies. The site includes links to two videos, an ACM paper on 3D photo arrangement, contact information for the team members, and a live demo. To run the demo, make sure you have the latest version of the Java Runtime Environment installed. For Mac users, if the demo does not launch, the site recommends viewing in the Camino browser.

The question, of course, is when will these technologies be ready for mass consumption? Though unsubstantiated, I have heard rumors that it takes a prohibitively long time for images to be organized and aligned. All of these previews are impressive and promise great utility for virtual globes and modeling the world. I am particularly taken by the idea of image tags propagating through the net, allowing an image of an unknown location to be identified.

Rumor mill and thanks

Rumor:
AGI might adopt NASA World Wind as its visualization engine.

Acknowledgements:

Also, referring to my previous entry about conferences, I'd like
to express my appreciation to the people that fed me, housed me,
and invited me to hang out with "the cool kids" over the last week.
These include:

Declan Butler
and Nature (for Sci Foo)
Chris DiBona and Google (for Sci Foo)
Tim O'Reilly and O' Reilly Publishing (for Sci Foo)
UCSB Geography and ESRI (for ESRI conference assistance)
Google Earth Team (for the geoblogger round up at the ESRI UC)
GeoMullah (for the geoblogger round up at the ESRI UC)

There were other important organizers of the Sci Foo conference that I did not meet. They worked hard and deserve credit, but I don't know all their names.

[Update: How could I forget one of Sci Foo's main organizers --> Timo Hannay of Nature!]

Spatial query processing utilizing Voronoi diagrams

I have not watched this video yet, but am posting it for my own notes. I thought others might be interested in it too.

Spatial query processing utilizing Voronoi diagrams (video)
Mehdi Sharifzadeh, USC Ph.D. Student, Presented at the Googleplex, August 10, 2006.

Conferences: ESRI and Sci Foo

While mostly I've been hunkered down coding, the brief pauses have been outstanding -- particularly the ESRI User Conference and Nature's Sci Foo.

ESRI User Conference
Other geobloggers have talked about the ESRI User Conference, but I'll add my two cents from a virtual globes perspective. First, I should disclose that I was only at the conference for two days, and did not attend any of the ArcGIS Explorer sessions. Mostly, I had conversations in hallways and chatted with exhibitors. Here's a few virtual globes-related tidbits:

  • The Sketchup guys were there, and they answered a few of my questions about Collada. Their online materials show examples of what Collada files look like and how a kmz can call one. Also, I probably knew this, but had forgotten: educators are eligible for free Google Sketchup licenses (link).

  • As I rounded the corner to enter my wife's talk on map design, I ran into John Hanke and some of the Google Earth team. I thanked them for sponsoring the geoblogger meetup. Of course, it's possible to buy a Google Earth t-shirt, but it just wears so much better when the Google Earth guys give one to you as a gift.

  • If you are into virtual globes and don't know about AGI STK, you should. I learned about their software several months ago, and enjoyed the opportunity to meet with one of the their engineers at the exhibit hall. AGI often serves as my muse when contemplating virtual globes: the software allows objects to interact with time, and the AGI add-on modules offer fairly sophisticated three-dimensional spatial analysis tools (ie. volumetric intersections, dynamic fields, video/terrain matching, etc). Visit their website, download their viewer, and check out their demos.




Nature/O'Reilly/Google Sci Foo
Immediately after the ESRI UC, I flew to San Jose, California, to attend Sci Foo, hosted by Nature, O'Reilly, and Google. With the invited participants, I knew the gathering would be outstanding. I attended talks on data visualization, data sharing (by the guys at Maya), the semantic web, virtual environments (Second Life), INSTEDD and early detection (Larry Brilliant), NASA World Wind, peer review, dynamic updating on the web, teaching the scientific process to students, and reforming patent law. Sci Foo presentations and discussions were self-organizing, and in the end, there were too many interesting things to talk about and not enough time. I heard many positive comments about the sessions I missed; for instance, people were buzzing about a demonstration of a three-dimensional video camera. Fortunately, there's a Google Video on the topic.

On Sunday morning, Ogle Earth's Stefan Geens led a geobrowsers session, and the discussion got lively enough that we ran out of time. In fact, Stefan didn't even get to finish his talk. He did however post his examples online. I'll post my talk as a Impress/Powerpoint file in the near future. Called "Big Maps! powers of ten, map of the cat, and digital earth," it's about design issues associated with integrating spatial analysis functionality at all scales.

[update: corrected a misspelling]

Thursday, August 03, 2006

Synchronizing GPS and photographs

Sony announced a GPS gadget that triggers from your camera's hot shoe. The device records location and time, to be later synchronized with your photograph. Here's a link to a story at Digital Photography Review. The device will go on sale next month for about $150.00.

Of course, if you have a logging GPS, you can do this already by synchronizing the timestamps on your photos with the time/location logs of your GPS. Here is a reasonable websearch (synchronize gps camera) to find the various methods to do just that.

Snooped out by Derrick Brashear via geowanking listserv (and posted on this blog so I can find it later).



Globes conference summary and announcement

A few weeks ago, Matt Nolan hosted a Virtual Globes Conference in Boulder, Colorado. Here is the link to Matt's summary, as well as a synopsis from participant Ron Schott.

Matt Nolan also sent me an email regarding several sessions on virtual globes at the December 11-15, 2006, American Geophysical Union meeting in San Francisco:

The Use of Internet-Based Virtual Globes in the Earth Sciences (Education Session);
Conveners: Nolan, Bailey
(details)

Geoscience Applications in Virtual Globes (Informatics Session); Conveners: Bailey, Dehn, Blair
(details)


The goals of these sessions include:

- Bringing together earth scientists, educators, and related individuals who are currently using or planning to use VGs in support of their work, to exchange ideas and insight. We are particularly interested in uses of VGs as natural hazards monitoring/mitigation tools, and for environmental concerns.

- The introduction of Virtual Globes to earth scientists unfamiliar with their abilities and demonstrate possible applications.

- Establishing links between users and developers of VGs to promote integration of the needs of the Earth Science community into these programs.

All presentations in the Informatics section will be given in using a novel form of poster presentation. These will take place in the poster display area, but AGU has agreed to set up a special area for our session, where presenters will be able to post a full-sized poster AND use a video projector to display their work on an adjacent 3'x4' space. High-speed internet and video projectors will be provided. The Educational session is expected to be a mixture of standard length talks and poster demonstrations in this same special poster venue. Note that we are not yet guaranteed an oral session and that AGU's conference guidelines allow for two first-author presentations as standard non-invited submissions if one is within an educational session.

Please contact any of the following with questions regarding Virtual Globes at AGU:

John Bailey, jbailey(at)arsc(dot)edu
Matt Nolan, matt.nolan(at)uaf(dot)edu
Jon Dehn, jdehn(at)gi(dot)alaska(dot)edu
Luke Blair, lblair(at)usgs(dot)gov

Vanished?

Okay, we haven't completely vanished. The three of us graduate students have been busy studying, reading, and writing. Karl is writing a manifesto on Digital Earth; Josh has been working on a paper related to infrastructure for virtual globes spatial analysis; and, I've been reading about mirror worlds, pondering nested coordinate systems, and making pushpins drive the streets of Santa Barbara within Google Earth. Anyway, the research group will be meeting Thursday to discuss the latest developments in Digital Earth (with Mike Goodchild and a visiting scholar). On Friday, I'm giving a talk to the CSISS SPACE Workshop on neogeography. After that, it's off to the ESRI User Conference for a day or two. I'll then be spending a few days with Ogle Earth's Stefan Geens and Nature's Declan Butler in Mountain View.