The extraction of buildings elevation information on UAV DSM

Compared with the traditional aerospace remote sensing platforms, The UAV was a new remote sensing platform, with the advantages of flexible, easy operation, high resolution and low-cost, which has been successfully applied to various field of the national economy. The manual extraction of buildings is an extremely labor-intensive, time-consuming work. So the automatic extraction of buildings elevation information has become the focus of international research. This paper take the UAV DSM data as the experimental data of the building´s elevation information extraction, adopt the method of regularization DSM data, firstly, based on the multi-scale morphological filtering method, Separate building elevation information from terrain elements, get the preliminary building elevation information. Then use the small region removal method combined morphological filtering to remove the small noise, get final building elevation information, Experiments show that this method can get a better result.