"Can't I already do 3D GIS for free?"
Every once in a while, someone asks us, "Can't I do this in one of the free virtual globes?" The answer is a little bit of yes and a whole lot of no.
Virtual Globes were designed for casual personal use, to be paid for through injected advertising. They are made to be fun, interesting tools for viewing hosted data, not for intensive professional use for exploring your own datasets. In this article we'll explore the three most common things we've seen people trying to accomplish with free virtual globes, the roadblocks they face, how Geoweb3d is different, and most importantly, why free can end up costing you more.
1) Exploring aerial photography
Ok, so everyone knows the first thing anyone does the first time they run a virtual globe is look at their house. Exploring aerial photography may be the most common use of virtual globes. Zipping around your city, looking at streets, parks, and rooftops, is fun and easy. But what about when you sweep a bit out of the city and things become more fuzzy? What if your town isn't close enough to a major city to be anything but a blob? What if you need to compare imagery from different years? That aerial photography may be on the webpage of your local GIS department (or maybe you are it), on the USGS Seamless Site, or elsewhere. You then have two options: overlays or Google Earth Enterprise. There are ways to create overlays from aerial photography, but they are restrictive, require labor (which is not free), and the end product is no longer a standard geospatial dataset that can be used for visualization and analysis in your GIS applications. This may be fine for small datasets, but any medium to large scale project will result in extensive labor costs and disappointing performance. Virtual globes make use of powerful server farms and clever caching mechanisms in order to provide interactive imagery performance. This relieves a heavy computational burden off of the end user's PC for viewing server hosted data. When viewing your own data, you're on your own, and depending on the scope your PC may not be able to handle it. To gain the same advantages that Google has, you need something like Google Enterprise, which comes with a hefty price tag.
For Geoweb3d users, life is much simpler. Click 'Load Imagery'. You just open your own files in their native formats, zoom to layer and start viewing. There is no need to prepare or reproject your data in any way. Instead of using caching or pre-rendering, Geoweb3d technology only loads what you need, when you need it, and only at the resolution your eye needs based on your location in 3d space. What that effectively means is you could load 1 foot resolution aerial photography for the entire globe and still enjoy the same interactive performance as with just your county. When you add up the labor, Geoweb3d may have just become cheaper than the 'free' products. Not quite? Keep reading.
2) Viewing detailed terrain data
When it comes to terrain, you have fewer options. Most free virtual globes do not allow you to add your own elevation data. Some users just settle for the global, limited resolution terrain provided out of the box. Others can't afford to. If your visualization project depends on detailed, accurate terrain, or requires alternative proposed or historic terrain models, once again you're on your own. The most common approach we've seen is to make a Sketchup model of the terrain. The problems with these workarounds are the same: no connection to the source data, limited interoperability, high labor costs, and performance issues for medium to large-scale projects.
The Geoweb3d approach, once again, is simple. Load your native data in its native format. Again, a large suite of supported GIS formats and on-the-fly reprojection mean no pre-processing is needed. To view alternative elevation models, be they proposed, historical, or archaeological, just load them all and turn them on and off with a simple checkbox. Want to load data of varying resolutions and worried about how to handle layer order? No problem, by default Geoweb3d will fuse all active elevation datasets on the fly, proving you with the highest resolution available for any location on the earth. If you require detailed terrain, Geoweb3d's low initial cost pays for itself, almost immediately and soon begins to save recurring costs when compared with 'free' products.
3) Viewing city models
Pop open your favorite virtual globe and go to a major city. You may very well see buildings start to fade in around you, each with detailed photographic textures. Once again, this is ideal for the home user, who will be thrilled to gaze down (or up) at a virtual Empire State Building. But for serious users that need control over their own virtual city, or for smaller communities that are not lucky enough to have their cities in 3D out of the box, the story is very diifferent. as with imagery, city models are cached and optimized server-side, releasing the end-user's PC of some serious heavy lifting. Users with their own city models learn very quickly that virtual globes are not the solution.
For example, on a recent site visit we were presented with a city model of over 2000 LIDAR-derived collada buildings with photographic textures. The entire city was broken up into several 'tiles'. We were told that only one or two tiles could be loaded into Google Earth at one time before it crashed or became unresponsive.. No application that had been tried was able to load more than one half of the city dataset at a time. We were put on the spot to see how many tiles Geoweb3d Desktop could handle. We loaded all seven, and then proceeded to add high-resolution imagery, LIDAR-derived terrain, and even lights and detailed trees. For a company that could spend up to two weeks patching together one fly-through video a tile at a time, this was significant to say the least. Another customer tested us on the micro scale, with an extremely detailed Revit model of a proposed construction project. The complete interior was modeled, down to each groove in the escalator steps to the pipes in the bathroom. Geoweb3d Desktop was able to provide an interactive walkthrough at an impressive framerate, while at the same time displaying a city-wide geographic backdrop. Geoweb3d can cut the complete visualization process for virtual cities and development planning down from weeks or months to minutes. For customers with accurate modeling requirements, Geoweb3d saves time and money when compared to free solutions.
Conclusion
Using advertising funded virtual globes for 3D GIS always involves workarounds, data cooking, and a reduced set of expectations. These applications were not developed for GIS but are commonly used due to the lack of alternatives and the perceieved cost savings. Basics features, such as layer reordering, integrating the attribute data, or displaying a dataset with hundreds of thousands or even millions of items, will soon frustrate the GIS professional. Like any application that gets used for something other than what it was designed and developed, the results will be limited and will diminish interoperability due to the need to restructure the original data for display purposes. Future efforts will be more aligned with social networking, for example, than those needs of the Web GIS user.
The end visualization product is almost always severed from the original data source. As the native source data changes, the same process to derive the data must be repeated again and again, resulting in long-term recurring labor costs. Managers are often deceived by the apparent capabilities of typical virtual globes and make the assumption that they can do the same things with their own data. The fact is that virtual globes are optimized for viewing their own cached, preprocessed data. Any serious visualization project that uses 'free' tools will suffer from reduced quality, capabilities, delayed production timelines, and alarmingly high labor costs. When compared to Geoweb3d, these 'free' applications become increasingly costly and limited in scope for the GIS professional.
