Value assessment of buildings based on machine learning

The PhD project exploits the potential of machine learning in combination with augmented reality as a tool to value assess buildings in order to create a better basis for decision making.

There are many cases, where developers decide to demolish historical buildings, because it is too costly and time consuming to value assess the building. This has a negative consequence for CO2 emission generally and wastes potential historical or cultural values related to the buildings. The utilization of ML and AR in relation to circular economy and adaptive reuse contains great potential, because of the complexity of the matter and the large amount of comparable data. This type of new tool will increase the amount of adaptive reuse and thereby decrease the waste of resources significantly and consequently generate new market opportunities and jobs within the circular economy.

It is the goal of this project to establish a digital workflow that generates comprehensive early phase models for the value assessment of existing buildings by (1) capturing and (2) analyzing a digital twin of the architecture, followed by a rich and thorough three-dimensional (3) representation of quantitative and qualitative data found in the digital twin. The research will include both quantitative datasets containing measurable data e.g., number of windows, CO2 emission factor, building materials, inventory etc. and qualitative data related to building-specific elements that holds value by virtue of their properties in relation to cultural heritage, historical narratives, local identity etc. The PhD project will research advancements in machine learning and augmented reality for detection and classification of constituted values in buildings. It will investigate how these technologies can enhance the complex processes related to adaptive reuse of the existing building stock. In order to arrive at a new model for a comprehensive value assessment, it will be examined which factors are the most important with regards to establishing a solid foundation for deciding whether adaptive reuse is a feasible solution in a specific case.