Fabric Engine: 3D LiDAR Rendering Via Web Browsers

Fabric Engine enables developers to create compute-intensive web applications with comparable performance to native, multi-threaded applications. Fabric Engine achieves this by tapping into all of the processing power of the user’s computer – something that normal web-based applications cannot do.  For the GIS market, this means that any web programmer can develop solutions for visualizing vast sets of geographic data such as LiDAR data in a web browser, something that represents a vast departure from how LiDAR data is currently managed and viewed.

Fabric Engine currently is only compatible with Firefox and Chrome browsers.  Users have to download and install the Fabric Engine plugin in order to run any of the demos that showcase Fabric Engine.  The company is working to have the Fabric Engine embedded so installing a plugin won’t be necessary.

There are quite a few demos to choose from, demonstrating Alembic IO, rendering, simulation, and bullet physics.  Being interested in the the geographic data visualization capabilities, I chose to take a look at a LiDAR scene of Mount St. Helen which demonstrated the LIDAR parser. The scene contains 5.6 million points and the data is parsed using a custom extension based on the libLAS library. Although not labeled, the controls allow for the adjustment of the shading options.  Zoom in, and the amount of detail this data set offers becomes apparent.  Panning, tilting, and zooming through the LiDAR data is fast and smooth.

The next scene is of Autzen Stadium in Eugene, Oregon and renders a colorized LiDAR data set that creates a remarkable facsimile of an imagery.  This scene is built from 10.5 million points.

Fabric Engine is developed by a startup of the same name.  Launched by founders Phil Taylor, Paul Doyle, and Peter Zion, the firm has a strong background in high-end 3D.

Videos explaining Fabric Engine can be viewed on Vimeo: http://vimeo.com/user6371427.  A list of live demos is here: http://demos.fabric-engine.com/.

So what’s next for Fabric Engine, especially as it applies to geographic capabilities?  Future plans include being able to create surfaces, editing the LiDAR data, and exporting the data.  The company is working with other developers who specialize in GIS and are working building their own implementations using Fabric Engine.  Paul Doyle also notes, “We recently built a volume renderer for a medical visualization demo. This rendering technique is useful when visualization geographic data as well. We also just started our Fabric Server alpha – this means people will be able to take their Fabric client applications and run them on the cloud. This is vital when dealing with massive data sets for two reasons: first, you don’t have the computation capability locally to work with huge data and second, if you did, you might not want to transfer all of that data over the wire i.e. a streamed result would be preferable.”

The company has also gotten Fabric Engine to work with Web Sockets which allows for a shared review of 3D scenes.  Web Sockets is a technology that allows people to collaborate and do things like shared review/editing on the same data. It has some interesting ramifications for remote collaboration. A demo can be viewed at http://vimeo.com/groups/fabric/videos/32201715.  In the demo you can see the other participants’ camera positions, and switch to their views.

The real niche expansion for GIS use potential will be client driven and primarily, Fabric Engine is a platform provider.  As Doyle states, “Everything on top of the core engine of Fabric is open-sourced, and Fabric is free for non-commercial use. So people have been taking our sample application and building their own tools on top of it.”  Fabric Engine’s open beta is available to developers by visiting fabric-engine.com/developers, the page also has developer support resources.

Share this article