This blog entry by Jeff Thurston, which describes the integration of the real world with the virtual, has echoed something that I often think about and try to show to people with video concepts because I just can't find the words to describe it. In Jeff's words:

While the article points out that imagery is now being collected horizontally, to enable realistic building facades and so on, one might very well argue that even that will not survive much longer as CCTV are now growing in number, covering just about every angle in most major cities. Perhaps real-time 3D at the street level is next up, in which case we will be walking through videos, rather than still images.

He's absolutely right, and I created a short video with the help of a clip from USC showing a simulated transition from satellite, to aerial, to 3D and finally to integrated "real" video stitched into the "virtual" world. Check it out.

Longer Example Of Fusing Video With Virtual Worlds

The MSNBC Rising from Ruin web site put together after hurricane Katrina used some virtual 360 degree pictures and videos to document some of the stories in Louisiana. They sent a team from Microsoft that was doing work in this area, and they performed the data capture and processing. The really neat thing about the video is that, while it is static in time, every user can stream it to their local system and pan the camera manually for a full 360 degree view, while simultaneously viewing their location on a Virtual Earth map.  

While anything is possible, there are many challenges to his concept of providing a real-time video walk-through,  not the least of which is the question of how the camera will move independently for every user of the system. I'm sure the answer will have something to do with quantum theory.

The German Government has released some 3D buildings in Google Earth. According to the site, approximately 10% of the city is modeled, amounting to about 44,000 buildings!

Following is a 2 minute clip of the experience from Washington, DC.

This is great news for the virtual earth community, as it proves that third party community contribution can really work. There are also some caveats to be learned from this model.

• The experience is very different from other Google-supported areas like Denver (video)
• There are multiple detail levels depending on the importance or architectural significance of the building in question
• Performance is dependent on where the models come from - in this case we assume a server in Berlin
• Rather than show all of the buildings at once, they have implemented their own "performance algorithm" that shows buildings in a narrow field of view

From the announcement:

In order to convey an impression of the cityscape of Berlin, some 800 facades across the city were photographed and the remaining building models textured - that is, assigned a facade - at random using these photos.

Personally, I don't want an "impression" of a cityscape. If I were a city planner, emergency manager or tax assessor, I would want to ensure that the building I was looking at is accurate both in geometry as well as texture, especially for those buildings that people don't look at every day. I also noticed that there is no "collision detection" on the buildings (i.e. as you get close to them, there are no obstructions, allowing you to inadvertently fly right through them.) And I assume it's a graphics driver glitch, by the buildings seemed to float about 300 feet above the ground on my system.

That aside, this is still a momentous achievement for which the team should be proud. It would be very interesting to know how many people contributed, how much it cost and how long this effort took, in order to gauge how long we'll have to wait for similar cities to be "born".

The Microsoft Virtual Earth team has automated the 3D building process with a pipeline that takes aerial imagery on one side, and through a photogrammetric process spits out highly accurate textured models on the other side in approximately 7 days (average 200 sq. km.) But I haven't yet seen a city with 44,000 buildings in it!

The nice thing about the Microsoft process is that you can be assured of the quality and performance of the buildings, including in highly dense urban areas such as this part of Boston's North End.

The image on the left is in Virtual Earth 3D, and the image to the right is in 2 dimensional Bird's Eye mode. Click on the images to see these areas in Virtual Earth.

This is all great progress on the road to creating a true 3D world. Great work Berlin!

I'm just catching up on blogs from the last couple of weeks and noticed Steve Lombardi's post on Virtual Earth 3D buildings video with Duncan Lawler of the MS Virtual Earth team.

So I was flying around Denver in 3D with the Road style turned on, and I began to notice some buildings way outside of the downtown area - like 15 miles south! That's one of the nice things about the Road style - with its light-colored background the buildings really pop. Make sure you turn on detail (SHIFT-D) or you'll just see lifeless gray boxes.

It struck me that I had no idea where all the textured buildings were in Virtual Earth, so I spent a few minutes this weekend exploring them. What I learned was that there are literally thousands (perhaps up to 100,000+) fully textured buildings in and around the initial batch of large cities.

Jerry Skaw linked to all of the cities, and I extended his collection with polygons that you can access and save as a Favorite here.

The roadmap at left shows the greater Denver area. The red pushpins are downtown Denver as shown in the example above. The rest of the area (20Miles x 5Miles) contains thousands of textured 3D buildings. In comparison, as of November 2006, the entire Second Life universe had about half as much territory as a single Virtual Earth city! Google Earth also has some nice textured buildings in downtown Denver, but they're limited to a small area neighborhood depicted by the pushpin.

My question of the day: when will television stations in the modeled areas start to use Virtual Earth in their broadcasts, instead of showing us lifeless gray buildings?