Papers
Enabling Component Reuse From Existing Buildings Through Machine Learning – Using Google Street View To Enhance Building Databases
Deepika Raghu ETH Zurich, Dept. of Civil, Environmental and Geomatic Engineering (D-BAUG), Stefano-Franscini-Platz 5, 8093 Zürich, Switzerland
Mathilde Marengo Institute for Advanced Architecture of Catalonia, Carrer de Pujades, 102, 08005, Barcelona, Spain
Areti Markopoulou Institute for Advanced Architecture of Catalonia, Carrer de Pujades, 102, 08005, Barcelona, Spain
Iacopo Neri Institute for Advanced Architecture of Catalonia, Carrer de Pujades, 102, 08005, Barcelona, Spain
Angelos Chronis Institute for Advanced Architecture of Catalonia, Carrer de Pujades, 102, 08005, Barcelona, Spain
Catherine De Wolf ETH Zurich, Dept. of Civil, Environmental and Geomatic Engineering (D-BAUG), Stefano-Franscini-Platz 5, 8093 Zürich, Switzerland
180 million tons of construction and demolition waste (CDW) is produced each year in the European Union (EU) and only 28% of it is recovered. The other 72% of CDW which is sent to landfills pose an environmental threat. Carbon emissions are rising not only due to the waste produced during construction but also due to increased raw material consumption. A very small proportion of these waste streams is reused or recycled despite the high recovery potential. There is a window of opportunity to respond to climate change by critically examining the possibility of reusing components from existing buildings. Most existing buildings do not have Building Information Modeling (BIM) models, making it difficult to carry out a whole-life assessment of the building to find salvageable components ahead of time. Scan-to-BIM is usually carried out just a few months before demolition and relies on expensive hardware and sophisticated software to process the data. Public databases also lack information on building characteristics required for component reuse. Hence, the challenge lies in expeditiously procuring information about reusable components from existing buildings. The study investigates the following question: what approaches and tools are needed to analyze the existing building stock and how can it be used to enable material and component reuse? The research is limited to the multifamily stock constructed between 1945 and 1975 as they are facing refurbishment or demolition needs. In this paper, a novel method to obtain data that can enable component reuse and be useful for building database enhancement is proposed. It effectively connects the digital world of Machine Learning (ML) with the physicality of architectural built space. Google Street View provides a large, accessible online database with images of buildings via the Google Street View API which is periodically updated. In conjunction with image processing techniques like deep learning, these images can be used to extract information that can complement existing cadastral data. 17788 ocular observations were conducted in Google Street View to analyze two building-specific characteristics: (1) facade material and (2) reusable components as such (window, doors and shutters) found on building facades in two cities: Barcelona and Zurich. Not all products are equally suitable for reuse and require an evaluation metric to understand which components can be reused effectively. Consequently, tailored reuse strategies that are defined by a priority order of waste prevention are put forth. ML shows promising potential to visually collect building-specific data of building characteristics that are relevant for component reuse. The data collected is used to create building level classification maps that can be used to define protocol and for urban planning. Such a classification map is mostly nonexistent and updating databases without automated methods can be very labor intensive. This research can be used to upscale limited information to regional or national levels. It can also be replicated in countries where available data about the existing building stock is insufficient.
Keywords: Machine Learning; Component Reuse; Google Street View; Material Banks; Building Database; Sdg11 Sustainable Cities And Communities; Sdg12 Responsible Consumption And Production