Remote Sensing-Based Approaches for Modeling 3D Vegetation Information in Urban Areas

While 3D information on cities has been frequently assessed in recent years, this is mainly limited to the built-up area. 3D vegetation information, for assessing urban ecosystem services such as carbon sequestration or cooling for example, is missing for most cities. New data and methods for calculating vegetation characteristics including remote sensing techniques offer large benefits for addressing this lack of information. The aim of this paper is to test different state-of-the art remote-sensing based techniques for modelling 3D vegetation in urban areas. For this purpose, we systematically compare different techniques for assessing vegetation parameters on two spatial scales: city scale and micro scale. In addition, we present results for the city-level scale for Berlin, Germany. First results show that we can derivate 3D tree vegetation parameters by using a forestry approach on high-resolution airborne LiDAR data (ALS) and multispectral Quick bird images with promising results. In comparison to field survey data we reached high accuracies concerning the derivation of tree location and number (detectable locations: all 70.18%, dominant 92.41 %) and got reliable information on tree height and crown width. In a next step, we will focus on additional types of urban vegetation, e.g. Shrubs on the city level and investigate the potentials of new remote sensing data with global coverage (Tandem-X) and unmanned airborne vehicles (UAV), which provide information on a micro scale.