Overview

The 3D engine has been built from the ground up leveraging the latest in video card GPU technology. It has been built exclusively for geospatial visualization with the ability to handle most native GIS formats. The spatial management and efficient rendering techniques allow for real time, game quality 3D GIS visualization.

The need to preprocess or restructure the data in preparation for efficient visualization is eliminated. The cost, delay, labor, and inaccuracies inherent in this approach are removed.  The data and the visualization are now seperated, allowing each to progress independently.  This is necessary for the visualization of both enterprise and dynamic data.

Inherent within this architecture is great scalability.  GPU technology is roughly doubling in performance every 6 months.  To put that in perspective, in 5 years we will have 1000 times the performance of todays technology.

Often, a stalemate develops between offline database formats and runtime visualization engines. Advancements cannot occur in one without effecting the other.  Compatability then becomes the overriding priority, bringing forward progress to a halt.

The most common 3D model formats can be loaded directly. Various web reuse initiatives are underway to index this data and make it easier to find and to share. This growth and ease of integration will continue to provide abundant 3D content for integration into the visualization.

The 3D visualization is network-aware and ready to be extended amongst multiple computers. Multichannel or immersive systems can be assembled on the fly to extend the field of view or represent datasets differently.

Attribute mapping is supported for communicating differences within each dataset. The 3D engine has extremely efficient spatial management and optimization, allowing for very large datasets to be manipulated and customized in real-time.

Additional 3D views can be created at runtime, efficiently sharing resources. Each can have their own settings such as time of day and view position. A user can cycle through these views and control each independently.

Once a dataset is loaded, it can be visualized and rendered in any number of ways. For example, a point shapefile can be loaded and represented as colored overlays, luminous lights, or 3D models, to name a few. The data and its representation are kept separate, allow the user to customize the visualization depending on the use case and needed emphasis.

To see Geoweb3d in action, please visit out gallery.

Loads most native GIS formats

Geoweb3d supports most common GIS, geospatial, and 3D model formats ranging from geodatabase to kml and Collada. Rich 3D visualizations are produced exclusively at runtime, directly from native formats. Absolutely no data restructuring or preprocessing is necessary prior to loading datasets into the 3D visualization.

The goal is to provide effortless integration with existing formats and interfaces that the Geoweb and the GIS industry depend on. This provides a level of interoperability that will ensure quick turnaround of visualization products and rapid sharing of data and information.  All data is reprojected on the fly, so no pre-processing of data is needed.

This visualization technology takes full advantage of the data loading, sharing, and analysis capabilities of both ArcGIS Engine and common open source APIs such as GDAL and OGR. Data from an endless variety of sources, such as the geodatabase, can be fused together into a captivating and highly informative scene, dynamically linked to the original source.

Supported vector formats include among others:

• Shapefile
• ESRI Personal Geodatabase
• ESRI File Geodatabase (w/ a valid ESRI license)
• ArcInfo Binary Coverage
• ArcInfo .e00 coverage
• GPS Exchange Format (GPX)
• MapInfo
• SDTS
• UK NTF
• GeoJSON
• GeoConcept

Supported raster formats include among others:

• GeoTIFF
• Jpeg2000
• ER Mapper Compressed Wavelet (ECW)
• USGS DOQ
• Erdas Imagine
• DTED
• National Imagery Transmission Format (NITF)
• Raster Product Format (RPF)
• Idrisi
• ESRI BIL
• VTP
• ArcInfo Binary Grid
• Leveller Heightfield
• USGS ASCII and SDTS DEM
• Terragen Terrain

Procedural scene generation

The recent advancements in video card technology allow for high fidelity, high performance, and highly accurate 3D geographic visualizations to be generated on the fly directly from the source data.

Graphics Processing Unit (GPU) technology has been fully leveraged within this architecture to manage extremely large and dense datasets while providing future scalability. Rich 3D visualizations requiring millions of polygons are now reality with adaptive scene generation capabilities providing for efficient real-time display.

Until recently, commodity hardware has been unable to provide the resources necessary to represent 3D geospatial data effectively. Forced to come up with a working solution, most scene generation solutions have had to restructure the data offline using proprietary tools producing proprietary formats.

Routinely, preprocessing of high fidelity datasets takes hours if not days and is required to prepare visualization suitable for the challenges of real-time display. The numerous proprietary and incompatible formats in use today have all but halted progress and collaboration within 3D scene generation. In these solutions, the link to the underlying source data is completely severed, making visualizations labor-intensive and immediately obsolete as soon as the source data is updated.

Once data is modified from its native and archived format, the costs and delays rise while the opportunity for effective reuse sharply declines. True interoperability becomes nearly impossible. The mandate for all future 3D visualization of geospatial data must be to use native and unaltered sources. Any other approach will continue to be seen as an implementation workaround.

Procedural scene generation is the only means to solve the problem. The technology acceleration of the video card industry, driven by the high-revenue gaming market, can now offer the hardware necessary to solve this problem efficiently and effectively.

Gaming-quality graphics

It has been the hope of both the defense and GIS visualization community that the gaming technology will adapt itself for geospatial 3D scene generation. This has not happened for many reasons and best left for a more detailed discussion.

Geoweb3d technology has been recently built from the ground up embracing all modern design, video card technology, and net centric data management. New rendering techniques can now present high quality, rich 3D scenes with modern detail such as realistic lighting. This can all be done while staying accurate to the source data resolution.

Many of these new rendering techniques have never been applied in the context of geospatial visualization. Geoweb3d can revolutionize the level of fidelity that can be achieved, well beyond what has been seen to date in existing virtual globes.

3D models

Most all standard 3D models formats can be loaded into the scene. Effective instancing, LODs, and other techniques are used to visualize very large quantities of high detailed 3D models within the scene.

Supported formats include 3D Studio (3ds), Collada, Openflight, DirectX, Wavefront OBJ, and Lightwave.  A wide variety of models can be downloaded from Google Warehouse and loaded directly into Geoweb3d.

Scalability

The demands for performance, realism, and scalability will continue to drive the pace of video card throughput, memory, and overall capability. Geoweb3d technology is at the forefront of all these GPU advancements. By procedurally generating scene content, the underlying architecture can continue to advance independent of the data. The fidelity, volume, speed, and overall quality will continue to improve with the coupling of software updates and video card advancements. Our strict adhenance to standards assures forward compatibility. Geoweb3d eliminates the need for intermediary data formats and binary compatibility.

Participants of our maintenance program will enjoy regular updates that will greatly extend the capabilities and performance of Geoweb3d, as we continue to adapt to support the latest advances in graphics hardware and Web GIS.

Attribute Mapping

By default, data representations within the 3D scene can be customized simply by clicking the representation in the table of contents and changing settings like color and elevation up on the shelf.  What if you need to communicate differences between categories or individual features in the dataset?

Geoweb3d Desktop offers an advanced attribute mapping interface that provides unprecedented flexibility while remaining easy to use.  One or more classifications can be defined in the specialized attribute mapping table of contents, where they can be customized on the shelf, just as if they were individual layers.

The power of Geoweb3d attribute mapping goes one step further.  The user is not limited to just mapping one attribute to one component of the visual representation.  Any number of attributes can be classified by any available aspect of the visualization.  For example, extruded building footprints could be given a height according to their real-world height, while specific types of buildings like police and fire stations could be color-coded.

Furthermore, the same visual aspect of the representation can be defined based on multiple attributes.  This is possible through Geoweb3d’s unique ‘cascaded attribute mapping’.  For anyone who has used cascading style sheets (CSS) in web design, this concept will already be familiar.  Any visual aspect of a representation can be defined multiple times, with the last applicable definition taking priority. 

For example, a point shapefile of a tree grove could be represented using tree models, with each tree type represented using an appropriate 3D model.  Dead trees, regardless of type, could be represented using a 3D model of a dead tree.  The former classification would be identified by mapping against a ‘tree type’ attribute while the latter would be identified by a ‘health’ attribute.  Through cascaded attribute mapping, the ‘health’ attribute map could take priority over the ‘tree type’ attribute map.  Therefore, a tree the is classified as both ‘pine’ and ‘healthy’ would appear as a live pine tree until its ‘health’ value was changed to ‘dead’, at which point it would appear as a dead tree.

Performance, quality, and accuracy

Leveraging GPU technology allows for realtime performance when navigating accurate, high quality 3D visualizations.

The performance is achieved through optimized software designed, architected, and developed specifically for real-time 3D GIS visualization.

The accuracy is maintained by means of the procedural, or adaptive, scene generation technique. Since all geometry is derived at runtime from the data, the terrain mesh necessary to hold truth to the data is runtime adjusted.

The quality of the 3D graphics is the foundation of our technology and will continually mature with time and technology.

Extend Amongst Multiple Computers

At the core of the Geoweb3d architecture is the ability to go beyond the computing resources of one computer. This can be used to extend the field of view or to render any number of alternative data representations simultaneously.

Establishing immersive applications is made easy and is scalable. One application can be the designate controller for all, assuring ease of use.

3D GIS visualizations can now escape the single view, single window application.

Data representations

A layer’s visual appearance is referred to in the application as its representation. A layer can have any number of these representations and view them simultaneously or individually. Groups can then be used to combine these layers and these representations which present an assortment of visualization options to communicate the geodata in rich 3D graphics.

A representation could be a thematic overlay or a realistic component of the 3D scene. Users have the ability to customize the appearance of the representation, such as color or height, globally or by attribute.  Each representation can be switched on or off instantly.

For example, a point shapefile containing all of a city's streetlamps can be simultaneously represented using an appropriate or exact 3D model and a light projection that will illuminate surrounding objects and terrain.  To turn off the lights, just disable the lightpoint representeation while leaving the 3D model representation active.  Similarly, a scene could be changed from summer to autumn simply by switching one group of representations off and another on.

Effectively communicate datasets

Users visualizing datasets in 3D are looking to communicate the result with others. This software goes beyond just realistic geospecific scene generation by allowing thematic representations of the datasets to be added and filtered at runtime. The result is a rich 3D scene that can be easily tailored to best communicate information and ideas.