Point clouds

Point clouds yield highly detailed information. For example, the second Dutch Elevation Dataset (AHN2) is so precise that even lampposts are visible. This degree of detail demands a large amount of storage and calculation. For this reason, researchers are working to develop several solutions using efficient data management.

The size of the AHN2 is gigantic; in all, it contains 640 billion data points. ‘In the future, we will have datasets containing tens of trillions of data points’, states Professor of GIS Technology Peter van Oosterom.
He shows an image of a round-about, resembling the Street View on Google Maps. ‘There are so many separate points that it is almost like a photograph. The point clouds are not easy to store in datasets, due to their size. It would be very convenient to be able to combine data. For example, the AHN2 could also show the owner of a building or select the high points along the route of a railway line’.

Van Oosterom and his colleague Oscar Martinez Rubi, on secondment from the Netherlands eScience Center, are investigating various solutions for improving accessibility. ‘For example, one solution involves the spatial clustering of data using so-called Morton or Hilbert code. We use a code with a value close to that of another code, which means that the associated points are also in close proximity to each other. This a a smart mathematical trick.’
The researchers also know that the transmission of data could be improved by transmitting the most characteristic points first. ‘For example, the highest and lowest points could be sent first, preventing the transmission of redundant data and giving the receiver an initial image of the surroundings’.

Van Oosterom and Martinez Rubi are also examining the desired and effective levels of detail. As anyone requesting information from the AHN2 can see, in 3D, everything that is close by is highly detailed, while objects that are farther away are much more vague. Van Oosterom is investigating whether it would be convenient to classify files according to their relative importance. For example, he assigns greater importance to the highest and lowest points. ‘With these and other adjustments, we hope to make the data much more accessible’.

Image: Fugro

Image: Fugro

Stay informed about the research

Receive the Delft Outlook newsletter 4 times a year