Sensor Networks and Smart Green Cities: Mapping Urban Tree Phenology

点击翻译按钮获取中文摘要

This PhD project aims to provide automatic insights into urban trees from ground level, high spatial and high temporal datasets. The project output will be designed to support the decision making, planning, and management of green infrastructure for sustainable smart green cities. Data from novel Sensor Networks and Internet of Things (IoT) devices such as CCTV cameras, dashcams, and traffic cameras will be obtained as sources of urban tree imagery. Computer vision techniques and deep learning algorithms will be applied to detect and segment urban trees from the data, and subsequently map the seasonal phenology of trees through obtaining metrics and indices including: greenness, leaf area, and vertical foliage profiles. Information on urban tree phenology is a valuable input variable in modelling carbon sequestration, urban hydrology, and the urban heat island effect. Phenology is also an important ecological indicator for understanding the feedback between climate change and vegetation productivity. Existing remote sensing and field observation approaches for monitoring phenology, in particular for trees in urban environments, have pitfalls in their resolutions which consequently limit our understanding of this natural phenomena. Urban trees play a key role in the functionality of city environments; ecosystem services have numerous benefits for cities and their citizens. This plethora of regulative, provisional, and cultural services to both society and the natural environment include: improving air quality, air temperature cooling, carbon sequestration, stormwater management, habitat provision, with urban trees are also thought to reduce stress and improve mental wellbeing. High resolution data on the phenology of urban trees will result in a better understanding of climate change impacts on tree productivity, and subsequently the actual value of ecosystem services. When also considering the potential to improve various environmental models, this project's outputs will be able to better inform the decision making and management of urban trees, such as for future planting plans, working towards a data driven, smart green city.