Detection and modelling of 3D trees from mobile laser scanning data

Abstract

Mobile laser scanning acquires massive point clouds in urban areas to provide high resolution data for 3D city modelling. A workflow for detecting and modelling trees from point clouds is presented. Emphasis lies on data reduction using an alpha shape approach. From the reduced point cloud the parameters are extracted to model the 3D trees using the Weber and Penn (1995) approach. The workflow is applied on two different sample data sets which were acquired with different mobile mapping systems and thus vary in quality and point density. The applied data reduction approach reduces the amount of data to process by about 95%. The tree models generated are consisting of a realistic trunk and branch structure of the tree crown. However, the inner branch structure of the tree crown is parameterised. Only the outer shape of the tree matches with the reality, which is sufficient for the requirements of visualization applications. For the tested areas the tree detection reaches a quality rate of 85 % and 78 % respectively. The comparison of the generated tree models against photographs and the original point cloud shows that the level of abstraction is sufficient for the integration of the tree models into 3D city models. 1.