The Gigapxl Project that is part of the newly updated default layer for Google 4.2 beta isn’t the only content out there that is taking advantage of the new KML 2.2 tags, specifically the <PhotoOverlay> tag, which lets you accurately position high resolution photograps and the viewer’s perspective so that the image fits in with the surrounding lanscape.

Gigapan.org, another gigapixel sharing project, has also georeferenced their photos and is now providing the KML network link for download.

Gigapan.org is the photo sharing site of the Global Connection Project at Carnegie Mellon University/NASA Ames Research Center, and Randy Sargent — who is a project scientist there — writes:

Anyone can upload a large panoramic image (50 megapixels minimum), click on “Place in Google Earth”, and you can adjust the panorama’s orientation and field of view until it’s correctly placed. gigapan.org does all the work to convert to something that’s viewable in Google Earth :-)

It looks like all this will soon be automated, at least for the dedicated amateur professional: there is a whole page dedicated to a soon-to-be released robotic mount for digital cameras that records the orientation and angle of the camera when a photo is taken. It looks suspiciously like the mount of my Celestron telescope:-) but that would certainly provide what’s needed to record the requisite data to feed a <PhotoOverlay> tag.

Some further panorama-themed notes:

1. The image zoom and movement controls are accessible to all content within <PhotoOverlay> tags; it is something everyone can use via KML, and not a Google-specific tool (as I had guessed this morning).
2. <PhotoOverlay> tags can deal with flat 2D images, partial or full cylinders, and also spheres.
3. The functionality is still buggy for me, running 4.2 beta on a Mac. Sometimes the larger image is upside down, sometimes it disappears, and quite often it doesn’t resolve when zooming in. I’m not sure if this is because of load on the servers, or if it will require an update. Is it working well for PCs? I should boot into Windows sometime… but life is so short!
4. Remember the much-promoted news from December 2006 where Pennsylvania announced it would use gigapan technology to make panoramas of US Civil War views? Well, now the technology is in place, so we can start expecting something to happen on that front…
5. So is gigapan.org like panoramio on steroids?:-)

It’s time for our quarterly pilgrimage to the Google Earth site to suckle at the download servers where a brand new version of Google Earth awaits for our edification: Google Earth 4.2 (beta).

I’ve just been playing a bit with 4.2, looking for new features:

Gigapxl photo layer: in the layers sidebar, a new item under Featured Content: Gigapixel photos from the Gigapxl Project:

It should be no real surprise that those photos have made their way into Google Earth, as Google Earth CTO Michael Jones is an avid gigapixelator. One thing to notice on the above popup is that there are links to other placemarks. Yes, indeed, that’s because Google Earth 4.2 has…

Support for KML 2.2: Time to go reread the KML 2.2 reference. You can now link from one placemark to another inside a KML file, including already loaded KML. Another thing the Gigapxl photo layer takes advantage of is the <PhotoOverlay> tag, which lets you orient a photo and the position of the viewer (among other things.) (The Gigapxl layer also comes with some new tools for zooming in on the photo itself, shown above — though I don’t think that’s part of KML 2.2.)

Google Sky: The most impressive and surprising new feature is also the one I almost missed: Can you find it?

Switch to Google Sky and you get the view from Earth of the Sky, using the same technique as Earth but inverted, so that you are staring from the center of a globe, and able to zoom in on a progressively higher resolution starscape. Or as Dave Bowman would say it, “My God, it’s full of stars!”

It’s all annotated too. Here’s what you can turn on in the suddenly morphed Layers sidepane:

It’s so new that it’s not without a few quirks: If you have some Earth-based placemarks turned on and switch to sky, you get this:

And the same goes for the Earth-based scale legend, which continues to show kilometers/miles. Also, there is no North Star, as the projection system suffers from the same problem we see at the poles on Earth:

But who’s quibbling? This is a wonderful new use of the Keyhole architecture.

If anyone wonders why Google might want to do this: I think the answer lies in the new search pane:

Google’s mission is no longer to organize the world’s information, but the universe’s. Vint Cerf’s talk at GeoWeb a few weeks ago now takes on a whole new salience. (He cautioned against developing a too Earth-centric internet.)

PS: We’re now going to have to start comparing Google Earth-Sky to other astronomy tools out there. Will Google shake up the astronomy field like it has GIS? We’ll have to wait for the official announcement to find out what the source of the high resolution sky imagery is.

(For the geeks: It’s now Google Earth 4.2.0180.1134 (beta) for Mac and Windows, and also Linux, Built on August 20, 2007)

[Update 10:12 UTC: The BBC has the scoop (Via Mapperz), interviewing Google’s Ed Parsons, who fills us in on some of the details:

“Click a button and the world flips round and you see the sky from that particular location,” explained Mr Parsons. “[The view] would be the constellations that you would see oriented in the sky on that particular day at that particular time.”

And the BBC writes:

Imagery for the system came from six research institutions including the Digital Sky Survey Consortium, the Palomar Observatory in California and the United Kingdom Astronomy Technology Centre.

Now Google Lat Long Blog also carries the announcement of what it calls “Sky in Google Earth” (not “Google Sky”, as the BBC calls it, which I suspect is what people will call it anyway.)]

[Edit: I got rid of the UTM mention i the original version of this post. That wasn’t new; my mistake. What is new is another option in the preferences pane: The ability to see coordinates as degrees with decimal minutes:

]

[Update 10:31 UTC: Astronomer Alberto Conti blogs about the genesis of Google Sky:

A few days later, I attended a National Virtual Observatory meeting together with Carol Christian. At the meeting we showed a KML that had been originally developed by STScI’s Frank Summers. Frank had taken the Hubble Press Releases and used the KML GroundOverlay structure to display them in Google Earth. It was the beginning of the realization that KML was versatile and that the Google GUI was malleable enough to be adapted to sky data! […]

At the meeting we met three people that were going to be the Google team that really made it all happen: Andrew Connolly, Ryan Scranton and Simon Krughoff. All three of them are astronomers, but they soon would become interns at Google and start working on GoogleSky.

I also see Alberto took the domain name oglesky.com for his blog. Fair enough. As long as nobody think’s we’re related:-)]

[Update 11:04 UTC: Reuters also has the story, and so does the New York Times.]

[Update 11:50: As Mapperz discovered, if you read the KML 2.2 reference, you’ll see that there is now mention of Flash support in Google Earth 4.2, so that you can embed YouTube videos. I haven’t been able to get this to work on my Mac with the latest Flash player installed. Has anyone else had any luck? This is a major and long-standing feature request, so it will be great to get this to work.]

That was easy: As of a few minutes ago, all georeferenced posts on IPY.org have a little embedded Google map besides them, showing you immediately what place the post is referring to. Before, we just linked through to a large Google Maps instance — we still do that, but embedding a map directly makes it all so much more playful.

(Link to post)

(Easy, because in the content management system we use we already have filled-in text fields for latitude, longitude, and a place name, so we just plugged those values into the <iframe> code as part of the template.)

I had to tweak the provided code a little to get my way: The left-hand column on IPY.org is only 240 pixels wide, so I squeezed the map to fit. Unfortunately, there is no way (that I know of) of just getting a placemark to show without the popup on an embedded map. The popup is too wide for this narrow map, so I had to opt for no popup and no placemark. Still, you can always click through to the big map, where the location is precisely marked.

Another little quirk I discovered with Google Maps: On Antarctica, you might have some very high resolution imagery and some very low resolution imagery, but nothing in between: If you hard-code the zoom level of your map, your view may fall between those two chairs, as happened on this post. Just keep on zooming in to see some very high resolution imagery of the Antarctic base in question, though you’d never know the imagery was there until you get to it.

That is is one way in which Google Maps’s image dataset differs from Google Earth’s. I don’t how big a technical challenge it would be, but it would be cool if embedded Google Maps could automatically guard against “over-zooming” of this kind, perhaps as an option you can turn on as a URL parameter. Another alternative: Filling in the gaps in those intermediate layers!

BTW, you can see all georeferenced IPY.org posts as KML here. That’s just another template using the coordinates to produce KML.

What are you doing reading this blog? Go to Google Lat Long Blog where you can read the news that Google Earth has now gotten a default traffic layer that shows congestion in US cities, a feature previously found only in Google Maps.

Google Lat Long blog also carries the very welcome news that Google Maps are now embeddable!!! Yes, you can now stick the URL of a KML file published to the web (or generated dynamically) into the Google Maps search window, get a map back, click on “Link to this map” and get the code to embed the map on a web page. You can even customize the size of the map. Google isn’t first with this feature, but considering that it is Google who wrote the API that let others embed their map, that’s a moot point:-)

Without further ado, then, here is the KML file generated by a Yahoo Pipe of my geotagged Flickr photos taken on my recent Alpine hike uploaded into Google Maps and embedded into this post:

View Larger Map

One immediate observation: Considering how easy it is to create such an embedded map, expect many more blogs to start adding maps to posts, ad hoc. I know I will. In fact, I’m itching to go redesign my personal blog right now.

Second: This is an inevitable improvement of the Google Maps service, but one which makes if more difficult for third-party mapping sites to differentiate themselves from Google’s in-house offering. It certainly keeps the pressure on third-party masher-uppers to keep innovating.

PS: Google sent via email a more detailed and more press release-ish announcement than what’s on the Google Lat Long Blog, and I’ve added it to the extended entry part of this post, below.

Continue reading Google announces embeddable Google Maps, traffic in Google Earth →

There is a new default layer in Google Earth, in the Featured Content folder, called Google Book Search. Turn it on and thousands of book icons suddenly populate the globe. These icons connect place names to their mention in books indexed by Google, providing yet another way to navigate information geospatially.

Does it work? Clearly yes, though in a rough-and-ready way. There is no doubt that in many places, linking different books by their mention of a specific place name creates wonderful serendipity, and this will keep us entertained for hours. But it is also a rather hit-or-miss affair: Some place names are unmistakable and unique — these will get accurate book references:

Esna, on the Nile.

But lots of other places (in Africa, for example) seem to attract Latin books or even Shakespearian English texts that have nothing to do with the place in question:

Argo, on the Nile

Elsewhere, shipping logs with ship names are mistakenly linked to place names, or similarly named places on the other side of the globe are referenced, or the link is due to mangled OCR-ing during scanning.

In other words, this layer suffers from the same difficulty that geo-parsers have when trying to turn newsfeeds into GeoRSS: Machines can’t (yet) extract sufficient context from a text to determine error-free when a word is a place name and when it isn’t — and if it is a place name, which place it refers to.

Is this a show stopper? Not at all. But there is one possible solution: Good old crowdsourcing. Why not have a little “wrong context” link in every book popup so that people can flag false positives? I know I would help along.

One more observation, though this is nothing that Google can do anything about: The books referenced are in the public domain, so they tend to be proceedings or annals or logs, or else older texts, often from the 19th century or earlier. This is interesting in its own right, but just don’t expect to find any references to Graham Greene novels. Another questions presents itself: How to deal with outdated names of places as mentioned in books? Sure, we all know Burma is Myanmar and Ceylon is Sri Lanka, but is there a time-line enabled register of all such places, and did Google use it? (Via Google Lat Long Blog)

The Navy Times reports in a long article dated Aug 19 that a maritime buff, Dan Twohig, found bird’s eye view imagery on Microsoft Virtual Earth showing the uncloaked propeller of a US Ohio class submarine in a dry dock. Dan posted about it to his website in early July, and also linked to the view:

These propellers designs are (were?) supposed to be a secret, and thus kept covered when out of the water. Navy Times interviews several people on the topic, and most sound fatalistic, except for a certain Norman Friedman. “a highly regarded authority and author on naval and military topics.” He blames everyone but the Navy for the mistake, if it is a mistake, and feels we should censor such imagery pre-emptively, just in case the Navy makes another mistake, or else the terrorists have already won:

“To make it easy for someone to get into a base like that is obscene. And that is something that can kill people. In huge numbers,” Friedman said. “Right now there are people out there in places like Waziristan who want us dead. They don’t have satellites, but they have wonderful fantasies. Why the hell make it easier for them?”

Because, you know, terrorists could really use a quieter submarine propeller. Elsewhere, he wonders “if the Navy has the temerity to go after Microsoft.”

Curiously, the Navy Times doesn’t reproduce the image of the propeller in the article, despite its self-described news value, nor does it link to Dan’s article, as netiquette would require. Isn’t that a bit the case of closing the barn door after the horse has bolted? And it took my just five minutes of Googling to find the original source, based on information in the article. Meanwhile, the Russians, Chinese and everyone else with a satellite likely has that same image by now, and at a much higher resolution.

But is this propeller even such a secret? I’d like to offer an alternate reading, based on the assumption that such “mistakes” are too big to be made accidentally. Just as the Chinese probably put their new Jin-class submarine on show for the rest of the world to see, to ensure everyone knows that it exists and can act as a deterrent (China’s military has obviously studied Dr. Strangelove), that particular propeller on the Ohio class submarine could be a ruse, out of date, or a signal to the rest of the world that the Ohio class is indeed dead silent, and thus a very capable deterrent. That would be another reason why US intelligence is so sanguine on the matter.

(In fact, a new comment on Dan’s original post suggests that such propellers are out of date, replaced by ducted pump jet propulsion on all the latest US, French and US models.)

One more thing: I was disappointed by Microsoft’s meek comment when asked about censorship by the Navy Times:

Asked about their policy on publishing such imagery, Microsoft officials offered a statement claiming that the company is willing to blur such imagery if asked.

That’s a lot more pliant that Google’s own robust defence of such imagery.

(I swear there is a Google Earth angle to this post. It’s just at the very end…)

Over at the Swedish Institute we’ve been experimenting with Second Life as a communications medium for engaging in public diplomacy, homing in on those specific functions that virtual worlds excel at, and for which there is no equivalent on the web. (We quickly found that if there is an equivalent on the web, it is likely much more efficient to use the web — there is no point trying to read any sizable amount of text in Second Life, for example.)

The main use we’re looking at is “social viewing” — where an in-world audience can watch live video on a screen at a specific location, simultaneously, giving them the opportunity to discuss the shared experience via chat. This is hard to replicate on the web: You can’t synchronize the viewing of YouTube videos with friends, for example, and even if you do all watch a live video clip — say a Steve Jobs keynote presentation — then there are still no easy ways to open a back-channel for commenting with like-minded strangers; if you use Skype, you only get your buddies. In Second Life, by contrast, everyone who is interested in a presentation just shows up at the appointed time and place. It’s a wonderful example of spatial navigation as an expression of interest — just as we do it in the real world.

With this technology, we plan to show lectures and Swedish short films, teach Swedish lessons, broadcast live Swedish jazz and DJ sessions — all we need is a laptop, a video camera, a microphone, a broadband connection, and Wirecast, a really cool live video editing tool (for Mac and Windows) that lets you create and send a very slick-looking video stream to a streaming server.

The whole system works flawlessly, and it lets us repurpose our content: An open session by the Nordic Council of Ministers, for example, might be broadcast to Second Life, a web page, and mobile phones. Only in Second Life do you get to engage in social viewing, though. In other situations, such as when a the director of a short film attends the Second Life screening to provide running commentary and takes questions from the audience, there is not much point in putting the video stream on the web — all the added value occurs in Second Life.

Where I am going with this on a blog about virtual globes? It soon became evident that this technology can serve very well as a teaching and outreach tool for a range of subjects. Wirecast comes with a helper application, Desktop Presenter, which lets you use the contents of desktop windows as an input to the broadcast, in addition to live video feeds and video, audio and still images from the hard drive. You also don’t need to broadcast live — you can record a performance to disk and publish it later to YouTube, just like a video podcast. I experimented a bit using Google Earth, and made this recording of me showing off some of the possibilities:

(Pardon the wonky first 5 seconds)

A couple of points about the above YouTube clip:

• All of this was done live, in one go, with me controlling both Wirecast and Google Earth from my laptop. A less stressful setup would be for somebody else to control Wirecast while the talking head just concentrates on talking.
• It was a very basic experimental set-up:

  

  If I had used a proper colored background for the chroma key, better lighting, a real DV camera and microphone (instead of what’s built into my MacBook Pro) then the result would have been much crisper. Also, Desktop Presenter can send screencasts to Wirecast from other computers on the LAN, so you could have Google Earth running on a second computer, for example, which would share the load. Alternatively, you could have a second monitor off the edge of the background screen, so that would-be meteorologists really can convincingly pretend to stare at what’s behind them.

• If you’re a schoolteacher, having students make their own broadcasts like this could be a wonderful way to get them interested in meteorology or geography.
• If you’re doing outreach for anything that involves a geospatial component (science, human rights, environmentalism), this would be a great way to show what you’re about to people who may not be familiar with KML, Google Earth or even computers.

There’s just one catch: cost. Wirecast isn’t cheap — $450, though there is a free demo — and if you want to send live video to more than a few recipients, you’ll need a special webhosting service that can run Quicktime Streaming Server (or the open-source equivalent). But if you’re a school and you want to use Wirecast for class projects, then the price certainly becomes reasonable.