A new point cloud processing pipeline has just come online at Voxxlr. In addition to filtering and sub-sampling, Voxxlr now also calculates the surface normals of individual points. This additional knowledge about the underlying topology makes it possible to visualize properties which are not readily apparent in the original point cloud. Below is a short illustration of these new analysis tools. Follow the links in each section to see a life demo.
Feature detection
A black and white illumination based on the pre-calculated point normals highlights subtle surface features that are undetectable in the original colored point cloud. For example, the image on the right clearly reveals the rafters underneath the roof.
Slope detection
Highlighting specific slopes also extracts information that is otherwise not immediately apparent. For example, the left hand image below shows all surfaces in the open pit mine with a slope below 8 degrees, effectively outlining the areas most equipment can safely access. Similarly, the image on the right highlights all roofs with an inclination between 25 and 45 degree.
Explore scan on the left
Explore scan on the right
Elevation levels
Although special post processing is not required to extract relative elevation from a point cloud, the normals are still useful in illuminating the scene. For example, the colored region in the image below highlights the coastline during high tide.
Explore this scan
Analyze your own point clouds by logging in at www.voxxlr.com and uploading data in e57, pts, ply, las or laz format. Voxxlr is capable of processing very large data sets using the vast computing resources of the Google Cloud infrastructure. Once a point cloud has been uploaded and processed, it becomes accessible via a standard web browser running the Voxxlr editor and 3D point cloud viewer. There is no need to install additional software and you are always working with the latest version.
Comments
Post a Comment