Assessing the potential of sub-metric resolution airborne data in the identification and characterization of urban impervious surfaces in Louvain-la-Neuve

Details

Supervisors

Defourny, Pierre ; Radoux, Julien

Faculty

Faculté des bioingénieurs

Degree label

Master [120] en sciences agronomiques et industries du vivant, à finalité spécialisée: Geo-Information Science and Earth Observation for Environmental Modelling and Management

Abstract

Urbanization has led to an increase in impervious surfaces such as roads, rooftops, and sidewalks, which prevent water from soaking into the ground. This can result in more surface runoff, higher flood risks, urban heat islands, and a loss of biodiversity. Accurate mapping and identification of these surfaces are essential for urban planning and management. This study explores the potential of using sub-metric resolution airborne data combined with LIDAR-derived height data masking to map and characterize urban impervious surfaces in Louvain-la-Neuve, Belgium. The research uses advanced texture features, including Haralick texture features and Structural Feature Set (SFS), as well as image saturation, to capture the distinct textures and colours of these surfaces. These features are processed using the Orfeo Toolbox (OTB) and classified with a random forest classifier. Although there were some challenges in distinguishing similar surface types, the study underscores the value of high-resolution data for urban surface mapping. The findings highlight the importance of differentiating between pervious, semi-impervious, and impervious surfaces for urban planning, such as in assessing urban heat island effects and planning green infrastructure. The thesis concludes that while the use of texture features on sub-metric resolution imagery is promising for detailed urban surface characterization, further research and refinement in classification methods are needed. Future research should aim to develop more extensive training datasets, incorporate additional data sources, and extend the geographical scope of study. These efforts will improve the generalizability and applicability of these techniques across various urban settings. The insights from this research are valuable for urban planning and environmental management, supporting sustainable urban development initiatives.