About the Project
Given the current forecasts in global population to reach 9.7 billion by 2050 , a huge increase in construction activities is expected to follow on the same trend. While boosting economic growth, activities such as construction and maintenance of buildings and infrastructures will also intensify the global demand for materials and energy, thus leading to an increase in depletion of finite natural resources, waste flow generation and a surge of the (already critical) levels of greenhouse gases. Significantly, buildings alone are currently responsible for about 39% of the total global energy-related atmospheric carbon emissions .
Given this trend, any feasible pathway to avoid catastrophic future scenarios can only be pursued if ways are found on how to decouple economic growth from current resource consumption levels. This can be achieved in practice by applying circular economy principles, that is to say, looking at the built environment as a primary reservoir of resources from which materials and products can be extracted, recycled and reused. To achieve this, assessment methods are required to characterise the building stock.
Built environment stocks research has reached increasing attention over the past decades , yet a strong methodological gap remains on how to achieve accuracy and reliability in a dynamic model at different scales and over time. Bottom-up approaches find applicability mostly at the small scale (i.e. city, neighbourhood) for which data/information with a high level of resolution can be feasibly achieved. On the contrary, for large scale analyses (e.g. at country or global level) top-down approaches, which are mostly underpinned by macro modelling of physical/socio-economic flows, are the only option currently available. Top down approaches however carry a high level of uncertainty mostly due to errors induced by the model aggregation and simplification. A third way to reconcile the advantages of both approaches is to leverage on the massive (and continuously growing) data availability of satellite and ground-level imagery. Existing studies for instance have already demonstrated the possibility for automated classification of building types solely based on their visual aspect .
The aim of this PhD project is therefore to advance the field by developing, implementing and applying machine-learning algorithms for automatic building stock characterisation. It is expected that these will be general enough to be globally applicable but the PhD should demonstrate its practical use by focusing on a specific region of the globe. Material quantities and intensities, building quantities and typologies, and aggregated reservoirs available at city/regional levels are all parts of the findings that this PhD is expected to produce in addition to the underpinning methodology.
The successful candidate will join the Resource Efficient Built Environment Lab (REBEL), a leading interdisciplinary research group focused on built environment sustainability.
A first degree (at least a 2.1) ideally in Computer Science (or Civil Engineering) with a good fundamental knowledge of machine learning/image classification.
English language requirement
IELTS score must be at least 6.5 (with no less than 6.0 in each of the four components. Other, equivalent qualifications will be accepted. Full details of the University’s policy are available online.
• Experience of fundamental knowledge in machine learning applications
• Competent in image recognition/classification
• Knowledge of programming languages
• Good written and oral communication skills
• Strong motivation, with evidence of independent research skills relevant to the project
• Good time management
Awareness of issues linked to sustainability in the built environment
For informal enquiries about this project, please contact Dr Bernardino D’Amico ([Email Address Removed])
 United Nations, Department of Economic and Social Affairs, Population Division (2019). World Population Prospects 2019: Ten Key Findings.
 Abergel, Thibaut, Brian Dean, and John Dulac. "Towards a zero-emission, efficient, and resilient buildings and construction sector: Global Status Report 2017." UN Environment and International Energy Agency: Paris, France (2017).
 Lanau, Maud, et al. "Taking stock of built environment stock studies: Progress and prospects." Environmental science & technology 53.15 (2019): 8499-8515.
 Kang, Jian, et al. "Building instance classification using street view images." ISPRS journal of photogrammetry and remote sensing 145 (2018): 44-59.