Historic digital survey: Conservation in the age of the fourth (digital) industrial revolution

It is well understood that digital technologies and applications are moving at considerable pace in society; such is the rise of ‘digital’ that it is accepted that we have entered the fourth industrial revolution, or Industry 4.0. The four revolutions are characterised as steam (18th century), factory production (late 19th century), computerisation (last third of the 20th century) and now cyber-physical.

Today, the third and fourth phases are considered transformational and pervade almost all sectors. Albeit slower than other sectors like manufacturing, the construction sector has also embraced innovation in attempts to unlock productivity and efficiency gains. The heritage sector, as a subset of the construction industry, could be argued to be even slower adopters of these technologies for a multitude of reasons; for example, conservation practices are often small and lacking resources both in terms of people and investment. This reduced diffusion and uptake of technologies arguably hinders productivity and efficiency gains in conservation projects which are all too often underfunded, especially in an age of financial austerity. Unlocking the potential that digital applications offer may result in tight project funds being better utilised and reallocated towards essential maintenance and fabric repair as opposed to spending it on laborious time-consuming activities which drain resources.

Traditional conservation practice was synonymous with analogue processes. These were often characterised as having limited interoperability (information sharing) and lacked the ability to automatically action other tasks that flowed from them. Hand-recorded survey drawings formed the basis of the historic documentation process and the development of subsequent project drawings. Conditions surveys, bills of quantities and an array of project documents relied upon these drawings to give contextualised accurate meaning to their specific outputs. Subsequent routine operations beyond major works were also analogue in nature and often limited in their use beyond the investigation of isolated building performance issues. For example, the use of glass or plastic calibrated ‘tell-tales’ for deriving data for the monitoring of progressive movement in buildings were manually measured, logged and interpreted.

Importantly, the information that was collected was often disparate in nature with limited connectivity to other forms of data-supporting analysis in other areas of building performance. Practically, the organisation of the physical documents also created issues, insomuch as these documents were classically held in manila folders in filing cabinets, or in drawing cabinets. Over time, these physical storage systems arguably increased the probability of loss of information associated with human error. The difficulties in being able to retrieve project documents effectively are compounded by the updating and replacement of information in a haphazard and incomplete manner.

Current practice for larger companies and projects may utilise digital applications that are characteristic of the era of the third industrial revolution. Today, on-site survey drawings are often replaced by digital reality capture acquisition (laser scanning and photogrammetry) and hand drawings such as elevations, sections and details are mainly created with CAD. These digital innovations have transformed the survey speed and enabled rapid file sharing in common work environments. These primary survey drawings form the basis of much of conservation practice, from recording and documentation of the ‘as built’ historic asset through to project visualisation and more. Survey acquisition has been assisted with the increasing and common place use of unmanned aerial vehicles (UAVs, aka drones). These have clear advantages in attaining survey data that is rich, cost effective (often obviating the need for scaffolds and working platforms) and, most importantly, safe.

Within the realm of survey and building operation, sensor technology for structural health monitoring and an array of performance monitoring of buildings are also becoming digitally enabled. Traditional methods have indeed become more sophisticated, such as the use of LVDTs for monitoring movement.

One of the benefits of digital data acquisition technologies is the ability to combine and overlay their data. For example, infrared thermography for hygrothermal performance analysis can be integrated onto laser scan data; indeed, some modern scanners can even be integrated with thermal imaging. While these technologies and applications are transformational, the data they provide is often not directly actionable without significant manual manipulation and processing with special, often complex, software. This lack of easy access to actionable information (direct actions that can be taken from the data thereby enhancing productivity) limits the potential and arguably hindered digital uptake as the commercial return on investment was limited beyond visualisation, documentation and recording.

An additional issue with classical digital technology deployment was a lack of objective guidance in the selection of the systems. For example, laser scanning captured the imagination of the public and practitioners alike. But, while laser scanning is appropriate in a number of contexts, alternative digital reality exists that may be cheaper, faster and result in better digital acquisition capabilities, such as photogrammetry.

Current so-called ‘digital’ approaches to information management, albeit facilitated by digital technologies, generally follow similar practice to classical physical document management, with digital documents (PDFs, for example) filed in electronic ‘manila’ folders on the computer, with little to no digital information integration or linkage. Increasing use of BIM is, however, noted in the wider construction sector which is often developed adopting a semantically rich digital model (linking digital information in a unified model). Software choices to author BIM models include ArchiCAD, REVIT, Bentley (Open Buildings Designer and Pro-Structure) or even open-source solutions like Blender. However, the importance of BIM transcends the creation of a model. One of the main reasons for the adoption of BIM is to create structured repositories of all information pertaining to the life of a building. These ‘digital twins’ can support more effective maintenance and conservation decision making. However, BIM repositories need not be complex and, in fact, do not even require 3D models. The heritage sector’s version of BIM is known as HBIM and reflects the increase in complexity in the creation of a model from an already constructed physical asset (as is). The process of creating a model, in this case, often starts with digital data from laser scans or photogrammetry, for example, which is then processed (mostly manually) to generate a BIM model, identifying and modelling the individual building elements. This process is commonly called ‘scan-to-BIM’. The model is semantically rich from this initial modelling step, but can be further enriched through use case - specific element properties which can store numbers, text or even links to other forms of digital data (see Sadeghineko et al, 2018).

For all of the perceived benefits of HBIM, its uptake is at present limited. It can be costly, requiring upskilling of the work force and wider practices and may be considered highly risky. In addition, the promise of efficiency can only be realised if the sector universally adopts the same working practices, which is a challenge due to sector fragmentation. That said, it is recognised that technological innovation may disrupt current working practices. The rate of new developments and harnessing of digital technologies will logically increase. The ready-availability of technologies such as handheld scanning and photogrammetry on smart phones is already transforming digital reality capture. This is particularly noteworthy for those in 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).

Figure 1: Semantic segmentation results of orthophotos of typical slated roof panels: the aim is for the programme to identify lead and slated areas and other features as distinct elements of the construction. (Reproduced from Lin et al, 2024).

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).

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 cost-effective 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

This article originally appeared in the Institute of Historic Building Conservation’s (IHBC’s) 2024 Yearbook. It was written by Frédéric Bosché and Alan Forster PhD. Frédéric Bosché PhD is a 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.

--Institute of Historic Building Conservation

Related articles on Designing Buildings Conservation.

• BIM.
• Conservation.
• BIM for heritage asset management.
• Digital modelling for retrofit.
• Digital techniques for heritage restoration.
• Digital technology in the historic environment.
• Drones.
• Heritage.
• Historic environment.
• IHBC articles.
• Institute of Historic Building Conservation.
• Listed buildings.
• Offsite construction.
• Modern methods of construction.
• Modular building.
• Using drones for inspection and survey.