In response to yesterday's post on 3D terrain a GEB reader in New Zealand, Paul van Dinther, forwarded me an article in a local paper which describes how North Shore City shared high resolution imagery and 3D terrain data with Google for Google Earth. The data was released back in April, and is quite impressive. I spent some time this morning flying around this nice looking suburb area north of Auckland. The aerial imagery is clear and sharp at about 6cm resolution, and the terrain is very well done. You can see shoreline cliffs, roads, and even footpaths in the sides of hills in the terrain.

I've made a brief GE 5 tour of the North Shore City area showing some of the high resolution aerial imagery, 3D terrain, the navy base, and views of Auckland. Here's a screenshot from North Head (a park on a hill) with an impressive view of Auckland and Mount Victoria on the right:

Many cities around the world have started promoting themselves by giving more accurate imagery, terrain, 3D models and other data to Google for use in their mapping tools. This is an excellent example of how Google Earth is getting better in a way that is mutually beneficial to users of Google Earth, and providing economic benefit to those living there.

One of the top stories yesterday was how NASA and Japan have released the data from their ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) satellite on the Internet - for free. What is significant about this is that the data covers 99% of the entire Earth's landmass terrain - verses 80% from the SRTM (Shuttle Radar Topography Mission) - which was the best previously available free public terrain data. ASTER's data is also higher resolution than SRTM which had a base resolution of 90m and at best 30m. ASTER's resolution has a base resolution of 30m, and can be higher under certain conditions (up to 7-10m - see documentation). The data released comes from over 1.3 million photos taken by the Aster satellite.

So, will Google Earth make use of this new ASTER data? The answer is: only a portion of the data. Google Earth already has terrain data in many areas that has higher resolution than the data provided by Aster (see below). However, ASTER provides data for remote locations (particularly high latitude regions) which previously wasn't readily available in these resolutions. And, Google Earth still has regions (such as remote pacific islands) which are using SRTM data (90m resolution). So, I'm sure Google will consider using the new data to improve the resolution in some regions. IF, they don't already have another source with better resolution. By the way, I asked Google about ASTER, and got the standard reply that they don't have anything to announce at this time (i.e. "no comment").

As far back as the February 2007, Google started introducing significantly higher resolution terrain. At that time they introduced 10m resolution terrain for the Swiss Alps, followed a few months later by 10m resolution terrain for the US and Canary Islands (read GEB review). In the more than two years since, Google has added millions of square kilometers of higher resolution terrain. Some of the terrain is much higher resolution. For example, the state of West Virginia has 3m resolution. Back in April of 2008, the city of Bergen, Norway got very high resolution data - which appeared to be based on LiDAR. Recently, Oslo, Norway also got very high resolution terrain data (also seemingly LiDAR) as well as 3D Buildings.

In recent months, Google has been regularly adding cities and countries with higher resolution terrain with nearly every imagery update (see for example the May 2009 update - new terrain is listed at the bottom). Google is no longer releasing details on the resolution of the data. Probably because the terrain resolution varies for each city. One drawback to higher resolution terrain is the data starts capturing things like buildings - not just the land - and it gets to be harder to tell the difference between good data and noise. But, the ability to see the terrain more accurately is well worth this risk in my opinion.

An important tip if you're interested in Google Earth's terrain data: Under the Google Earth Options is a slider called "Terrain Quality". If you push that slider to the right, you can see higher resolution terrain. A word of caution: higher quality terrain means more 3D data - which can slow down your update speeds in Google Earth. After viewing higher resolution terrain, I recommend putting the slider back to closer to the middle for every day use. If you have a faster computer with a powerful video card, you might be able to get away with a higher setting. (Read more tips about optimizing GE performance).

NASA is conducting research from their Ames Research Center in California with a robot called "K10", located in northern Arizona, to simulate tele-operations for exploring the moon. They are using Google Earth extensively for both planning, visualization, and operations. In other words, Google Earth is a visualization tool for telepresence. NASA uses a black and white higher res image of their target area overlayed in GE (maybe the black and white makes it look more like the moon?). They look at the imagery to determine reconnaissance objectives and plan routes. They have developed a planning tool that works interactively with Google Earth (using network links and overlays) to develop and assign tasks for the robot and put placemarks and routes on the map. Then the task plans are sent to the robot. The robot sends back photos to NASA which are also placed in placemarks in Google Earth, and a track of the robots path is also updated continously. You can read more details and see more screenshots at the Robotic Recon web site. I'd like to get my hands on a sample KML file to see what it looks like. There is a large team of engineers involved in the Robotic Recon project, including folks from different NASA centers and several universities. And, the Robotic Recon blog shows lots of screenshots of Google Earth in use.

By the way, this story is particularly interesting to me because more than 15 years ago I developed similar tele-presence simulation software at NASA Johnson Space Center to visualize remote-operations over the Internet using a Silicon Graphics workstations. I developed the 3D software for the visualization parts, and the Internet client/server communications. In 1992, while at International Space University in Japan, I used a workstation to control a robot arm NASA had back in Houston over an Internet connection using a crude frame-grab camera (single-frame webcam) to verify position. This is so déjà vu! Part of the reason I was instantly attracted to Google Earth when it first came out, was its similarity to the software I had written. Full circle.