Tuesday, 28 June 2011

Automated 3D Digital Landscapes

Together with the digital mapping the 3D modes have become very popular and are of most mapping services these days. There is the basic terrain that can be rendered in 3D, but also the buildings.

Google had from very early on 3D buildings for cities, gradually expanding on numbers and quality. They started with grey volumes and now show good quality photo mapped buildings. To get here Google tried crowd sourcing the work in different ways and were really successful. They offer an online tools, the Building Maker, for people to use together with the required data such as location, aerial imagery and images of the building or texture mapping. This was back in 2009.

SenseFly UAV
Image taken from engadget / 3D model of an urban fragment generated from areal images recorded by an SenseFy UAV.

Other companies tried different technologies. Yell Maps were one of the first to show full coverage in 3D for cities using 3D models built from satellite imagery. It is based on lidar scans that created the basic mesh for the topography and then automatically mappen on with images.

This is a very different approach to the Google model because basically the buildings and the topography are mapped out at once. It is a high res topography scan that will include the buildings. This is sort of what other companies are using now for the 3D visualisation of online maps as for example the Nokia OVI maps or the maps available on the Swedish search engine hitta.se.

lidar scan
Image taken from searchmesh / The principles of a lidar scan.

At EPFL in Lausanne a team of scientists has extended on this research and developed a drone based mapping version of a similar technology. It is however, no longer based on laser scanning of the terrain for the point mesh, but instead they are using the images themselves to create the point mesh.

The drone is something like the SenseFly. An UAV works autonomous but can be controlled in rel time and captured images re available immediately after landing. See a clip HERE and a post on Digital Urban HERE.

SenseFly UAV
Image taken from geeky-gadgets / The UAV developed by SenseFly. It comes with a 12m camera and controle software in a neat flatpack box.

The drone will cover an area multiple times taking pictures from different angles as it passes overhead. The computer will match all images calculating the differences and produce the mesh in amazing detail.

This sort of brings the 3D modeling back to the crowd where everybody can join in and produce parts of the digital environment. On a larger scale this would also provide time based imagery with links to a great archive. It could become a tool for project like the Grassroot Mapping project (discussed on uT HERE) and similar projects of public involvement and feed into the Community Remote Sensing CRS movement.

Probably the commercial version is too expensive at $10'000 and mainly focusing on faming applications and land surveys. But maybe an adapted version using an iPhone and the photostich software to merge the images.

Via sunFoundation

1 comment:

Jesse McGraw said...

And maps services just keep on stepping it up, giving people the most accurate map terrains. It's pretty amusing to see 3D landscapes - it's as if you're taken exactly to that specific place. Anyone with a smartphone can pretty much type in a location and see how interesting it looks like on the map. From land surveyors to casual users, there's always something to see on the 3D map.