Extraction of Urban Environmental Quality Indicators using LiDAR-Based Digital Surface Models

The visualization of specific 3-D urban scenes can be done calling upon different techniques, from those more traditional, such as photogrammetry, to the most advanced ones, such as laser scanning that uses different techniques and algorithms of selection and modelling of 3-D point clouds. The use and utility of this kind of data for the study of urban development remain however debatable. Indeed, indicators for urban development and durability are highly necessary and the best methodology to build them is largely open. This thesis anticipates the use of 2-D and 3-D models and data for the environmental analysis of cities, aiming to provide useful tools for urban planning and design. According to end-users requirements, the extraction of urban environmental quality (UEQ) indicators from 2-D and 3-D information using innovative methods is proposed and implemented, which is based on recent research on computational algorithms for the analysis, evaluation, management and design of the urban space. Moreover, results that can be obtained with different data sources and aggregation methods are compared. In particular, the main advantages of urban models generated from LiDAR data are highlighted. In consequence, an iterative process is proposed, involving professionals of various fields, aiming at improving the utility of those indicators for the support of applied decision activities related to the sustainable development of cities. This process is sub-divided in three correlated steps: A preliminary inquiry concerning the user requirements for the implementation of a 3-D project of the State/City of Geneva was launched. Based on the obtained replies, several potential applications related to both the definition and extraction of urban indicators were identified, and also, end-users were classified into 6 different domains: 1– architecture, urbanism and territory planning; 2– urban traffic (motor vehicles, trains and airplanes); 3– environment and energy; 4– pedestrian and cyclist mobility; 5– security and emergency situations management; 6– underground information; Based on point 1. and according to the assessment of the specific needs among each of these domains, several interviews were carried out in which 25 end-users decided to focus on UEQ indicators considering three main stakes: 1– assessment of the morphological properties of the urban texture; 2– exploration of the solar potential on the urban fabric; 3– estimation of the energy demand on the urban fabric. Many empirical case-studies are emphasized, mostly for the city of Geneva, and also for the cities of Lausanne and Florence. These indicators are extracted from the segmentation of planar roof areas using classified LiDAR point clouds and the use of image processing techniques based on Digital Elevation Models (DEM) and Digital Height Models (DHM), defined in this thesis as 2.5-Digital Urban Surface Models (2.5-DUSM) and normalized 2.5-Digital Urban Surface Models (n2.5-DUSM) respectively.