44 YEARBOOK 2024 the sector that are digital natives. Easy acquisition of data, combined with progressive development of algorithms that effectively ‘make the data speak’ will ease the generation of actionable information. Examples that are currently noted are automatic segmentation of reality capture data to distinguish stone from mortar regions in masonry, and leadwork from slates in roofs (see Figure 1), and machine learning algorithms for automatic defect detection in masonry or roofs (see Figure 2). The use of digital applications such as BIM modelling in conservation practice will logically increase as the digital literacy of staff grows. The promise of meaningful automated scanto-BIM appears someway off, although the extraordinary pace of development of AI may address this challenge sooner than we can conceive. Moving beyond the creation of a BIM model, developments in ‘digital twin’ promise much for the heritage sector, especially in the real-time data collection on an array of performance related factors, and for the structured recording of conservation data and information over time. These applications could be transformational for historic buildings and their collections. For example, the real-time monitoring of hygrothermal response to changing environmental conditions (intense external rainfall or high relative humidly) that is fed back to the digital model could flag the need for intervention to proactively prevent damage. Other applications of digital twinning relate to structural health monitoring with embedded micro-sensors or MEMS (micro electromechanical systems) which are capable of relating data on movement in the building fabric (such as rotation, leaning, deformation, moisture or thermal movement). Again, the data helps in understanding the structural performance which requires longer term input, such as seasonal data on moisture contents of clay rich soils and whether movement and associated fracture patterns are static or dynamic. Robotics in the main construction sector is starting to gain traction, especially in modular construction processes. Heritage is an inherently more challenging proposition due to the complexity of the architectural arrangements and geometries, and the variability of the materials. Promise is noted in robotic platforms that enable accurate deployment of sensors for close range monitoring and even physical contact with the building fabric (for example, penetrating radar or micro drilling). That said, climbing and soft robotics for intervention appear a long way off and are again, arguably, the most difficult to deliver due to the bespoke nature of historic buildings and their inherent complexity. Advances in CNC cutting is already gaining traction. The ability to accurately reproduce carved enrichments is becoming easier, with connectivity between scanned point cloud data combined with advances in cutting technologies. In addition, digital printing technologies are also becoming prevalent with an array of printable materials being noted. Of particular interest is the use of printable mortars to create large scale architectural elements. These techniques clearly offer costeffective possibilities for an array of conservation applications, but they also throw up philosophical concerns relating to authenticity, architectural legibility and a reduction or loss of traditional craft skills. The wider construction sector faces growing competition from offsite modular construction in particular (a development which may, in fact, help address the sector’s chronic skills shortage), and there are legitimate concerns for job security arising from robotics and AI. However, conservators and heritage specialists are arguably safe for the foreseeable future, given the bespoke nature and complexity of almost all traditional structures. RECOMMENDED READING Cyberbuild, University of Edinburgh, cyberbuild.eng.ed.ac.uk Idjaton K, Janvier R, Balawi M, Desquesnes X, Brunetaud X, Treuillet S, “Detection of limestone spalling in 3D survey images using deep learning”, Automation in Construction, 152, article 104919, 2023 Ross P & Maynard K, “Towards a 4th industrial revolution”, Intelligent Buildings International, 13, 159–161, 2021 Valero E, Bosché F, Forster A.M, Hyslop E, “Historic Digital Survey: Reality capture and automatic data processing for the interpretation and analysis of historic architectural rubble masonry”. Proceedings of the 11th International Conference on Structural Analysis of Historical Constructions, 2018 Frédéric Bosché PhD is Reader in Construction Informatics at the University of Edinburgh, and Past President of the International Association for Automation and Robotics in Construction. Alan Forster PhD is Associate Professor in building conservation, low carbon materials and architectural technology at Heriot-Watt University, Edinburgh. Figure 2: Example of algorithms, now commonly based on Deep Neural Networks, for detecting defects in masonry. (Reproduced with permission from Idjaton et al, 2023).
RkJQdWJsaXNoZXIy MjgyMjA=