Acta Univ. Agric. Silvic. Mendelianae Brun. 2013, 61, 2415-2421

https://doi.org/10.11118/actaun201361072415
Published online 2013-12-24

Point cloud processing for smart systems

Jaromír Landa, David Procházka, Jiří Šťastný

Department of Informatics, Mendel University in Brno, Zemědělská 1, 613 00, Brno, Czech Republic

High population as well as the economical tension emphasises the necessity of effective city management – from land use planning to urban green maintenance. The management effectiveness is based on precise knowledge of the city environment. Point clouds generated by mobile and terrestrial laser scanners provide precise data about objects in the scanner vicinity. From these data pieces the state of the roads, buildings, trees and other objects important for this decision-making process can be obtained. Generally, they can support the idea of “smart” or at least “smarter” cities.
Unfortunately the point clouds do not provide this type of information automatically. It has to be extracted. This extraction is done by expert personnel or by object recognition software. As the point clouds can represent large areas (streets or even cities), usage of expert personnel to identify the required objects can be very time-consuming, therefore cost ineffective. Object recognition software allows us to detect and identify required objects semi-automatically or automatically.
The first part of the article reviews and analyses the state of current art point cloud object recognition techniques. The following part presents common formats used for point cloud storage and frequently used software tools for point cloud processing. Further, a method for extraction of geospatial information about detected objects is proposed. Therefore, the method can be used not only to recognize the existence and shape of certain objects, but also to retrieve their geospatial properties. These objects can be later directly used in various GIS systems for further analyses.

References

20 live references