Automatic Extraction of Tree Rows and Hedges by Data Integration

This paper mainly focuses on the integration of GIS data, DSM information and CIR stereo imagery to derive automatically tree rows and hedges in the open landscape. Different approaches, such as segmentation by CIE L*a*b, edge extraction, linking and grouping, and verifying with 3D information, are successfully integrated to extract the objects of interest. The extracted tree rows and hedges can be used to update the GIS database, to derive wind erosion risk fields for soil monitoring and preservation.