Monitoring of the Urban Environment
Improving our understanding of urban ecosystems through remote sensing
Sustainable management of urban areas is crucial to urban biodiversity and to the quality of life of present and future generations of urban citizens. The transition towards more sustainable urban futures requires the development of new, transdisciplinary research approaches contributing to a better understanding of the complex interactions between social, natural and human-engineered systems. Remote sensing is playing an increasingly important role in improving our knowledge on the functioning of urban ecosystems, on the biophysical changes that take place in urban environments, and how these changes affect both humans and nature.
Different sensor technologies (multispectral and hyperspectral imaging, thermal remote sensing, airborne and terrestrial laser scanning, …), have made it possible to monitor land cover in urban environments at multiple scales, from the neighborhood or street block level up to the scale of individual objects such as trees or buildings. At CGIS we combine different types of sensor data to assess (bio)physical and structural characteristics of urban spaces and of the grey and green infrastructure that makes up these spaces.
Sealed surface mapping
Urban expansion and urban densification result in massive land cover change, with open spaces being substituted with grey infrastructure, consisting of impervious or semi-impervious man-made materials. Over the years CGIS has developed and tested alternative approaches for mapping sealed surface cover from remotely sensed imagery, using or combining data from sensors operating at different spatial and spectral resolutions.
One line of research focuses on developing approaches for mapping sealed surface cover fractions from medium resolution remote sensing data (type Landsat, Sentinel-2) using machine learning based methods. Recent work on this topic is looking into issues of transferability by exploring the possibilities of developing generalized models for mapping sealed surface fractions over multiple cities using multi-site urban spectral libraries for model training. Next to mapping sealed surface cover at medium resolution our team also has experience in detailed mapping of roof and road construction materials from airborne hyperspectral data.
The sealed surface cover maps which we produce have been used in studies on monitoring of urban growth, modelling of urban growth dynamics, and in assessing impacts of sealed surface cover on runoff and ground water recharge.
Mapping of green infrastructure
Green infrastructure in cities provides a whole range of ecosystem services contributing to the well-being of urban citizens. Improving the connectivity of urban green spaces is crucial for preserving local biodiversity in urban areas. Over the last decade there has been a growing interest, both from researchers and local governance, in assessing the ecosystem services provided by urban green. Urban trees in particular play an important role in the supply of ecosystem services and in safeguarding biological diversity in cities.
The ecosystem service supply delivered by individual trees mainly depends on the type of tree (genus or species) and its structural characteristics (tree crown volume, crown density, leaf type). CGIS conducts research on the development of urban tree inventories based on fusion of different remote sensing technologies that provide information on the phenology, crown volume and crown structure of individual trees. Deep learning models are being developed to detect and identify individual trees at genus/species level.
The outcome of this work is used in studies on the assessment of ecosystem services and disservices provided by different types of trees and on the role of urban trees as food source or habitat for urban fauna.
