Semi-automatic Analysis of Aerial Photos for the Detection of Settlement and Traffic Areas for a Sustainable Land Management.
This paper deals with the semi-automatic analysis of aerial photos in urban areas. Germany is planning to decrease daily land consumption to 30 ha per day by the year 2020. To achieve this objective a qualitative sustainable land management planning is necessary incorporating urban, social, economical and ecological aspects. Based on existing data of the city of Osnabrück land cover information shall be derived in timely and cost-effective manner for urban and regional planning that can be incorporated into the existing planning methods. For this an area-wide object-based urban classification is performed in two steps. It is supported by vector-based extraction and smoothing methods. Firstly the initial
segmentation is supported by an algorithm for the detection of hipped roofs. Afterwards classified object geometries are being improved
through smoothing methods. In the second step all object classes are being classified and smoothed again using automatic real estate map
data (ALK). The aim is to reduce manual postprocessing efforts.
Photogrammetrie - Fernerkundung - Geoinformation 01/2008; · 0.38 Impact Factor