Digital Terrain model reconstruction in urban areas from airborne laser scanning data: the method and an example for Pavia (Northen Italy)

Light detection and ranging (LIDAR) techniques represent a new and fruitful approach in the determination of digital surface models. One of the goals in processing this data is to set up filtering methods which automatically allows to extract the ground and the features (buildings, vegetation, etc.) superimposed on the terrain itself. In our work the emphasis is focused on the first topic. The implemented method takes advantage of the use of spline functions regularised by means of Tychonov functional in a least-squares approach. Firstly, the DSM pixels have been classified in order to previously detect any edges of the non-terrain feature. Then all the pixels corresponding to the ground, by means of a region growing algorithm, has been identified through a correction procedure. Finally, by a new interpolation on the classified ground pixel only, we can derive the digital terrain model. In the paper the processing methodology is discussed; and a first extensive example is presented.