A system to automatically classify LIDAR for use within RF propagation modelling

Many technologies and applications now necessitate an awareness of their geographical surroundings, typically employing an array of sensors to capture the environment. A key application is telecommunication network planning which benefits from the utilisation of RF propagation tools which incorporate representations of target environments typically sourced from high resolution aerial photography and/or LIDAR point clouds. However, the amount of data associated with LIDAR scanning can be very large, permutation invariant and clustered. Manually classifying this data, to maximise its utility in a propagation model, is not easily scaleable; being both labour intensive and time consuming. This paper describes a system which facilitates the automatic classification of point cloud data and its subsequent translation as wireframe meshes into a propagation model. Testing of automatically classified versus hand-labelled clutter results in comparable performance, with the average difference across all measurements of the automated approach outperforming hand-labelled data by circa 2.5 dB.