Energy efficiency evaluations of buildings using VHR thermal data

 Suggested by: Dirk Tiede, Felix Kröber

Short description: 

Thermal imagery acquired by satellite-borne sensors such as MODIS, Landsat, Sentinel-3 and ECOSTRESS have been limited to a spatial resolution of 70 m or more per pixel. While this enables to analyse city-scale effects such as urban heat islands, analyses on the level of individual buildings are unfeasible with this resolution. However, recent technological advancements bringing the resolution of satellite-borne images down to 3.5 m could make such analyses possible for the first time. The application domain of energy efficiency evaluations of buildings is of particular interest given their significant share of energy consumption (40% EU-wide) and their contribution to greenhouse gas emissions (36% EU-wide). 

The aim of this thesis is to develop a model based on thermal remote sensing data for assessing building properties, with a particular emphasis on energy efficiencies. A multiscale approach is suggested analysing model predictions made on different spatial levels from districts via neighbourhoods down to individual buildings. The research is grounded on a novel Very High Resolution (VHR) thermal dataset that spans approximately 75km2 in Liverpool, UK. For the same area, we have point-wise energy performance data available for over 75,000 houses. Methodologically, statistical models of different complexity ranging from straightforward regression models to advanced machine learning techniques should be tested. Optionally, emphasis on spatially explicit regression models, including spatial lag models and considerations for spatial autocorrelation would be desirable. 

Suggested reading:

Pasichnyi, O., Wallin, J., Levihn, F., Shahrokni, H., & Kordas, O. (2019). Energy performance certificates—New opportunities for data-enabled urban energy policy instruments? Energy Policy, 127, 486–499. https://doi.org/10.1016/j.enpol.2018.11.051 

Dennett, A. M. and A. 2023. Chapter 9 GWR and spatially lagged regression | CASA0005 Geographic Information Systems and Science. Retrieved 30 October 2023, from https://andrewmaclachlan.github.io/CASA0005repo/ 

Start/finish: ASAP

Prerequisites/qualifications: 

• Understanding of remote sensing imagery analysis 
• Methodological knowledge in the field of statistics incl. spatial statistics 
• Basic programming skills (in Python or R)