Laser scanning intensity fingerprint: 3D visualisation and analysis of building surface deficiencies

Laser Scanning Intensity Fingerprint: 3D Visualisation and Analysis of Building Surface Deficiencies Daniel Antón, Manuel J. Carretero-Ayuso, Juan Moyano-Campos, and Juan E. Nieto-Julián Abstract Terrestrial laser scanning (TLS) is a widely used technology in numerous sectors since it enables the recording of both geometric data and colour information of the objects. Moreover, this remote sensing technique allows for producing point clouds enhanced with the reflection intensity of the laser beam. Scientific research has used those data to detect and assess building surface deficiencies. However, the laser scanning intensity fingerprint of a building pathology is yet to be addressed. Thus, this research quantitatively analyses the distribution of point cloud intensities throughout the object geometry to show changes against the general context of the building component surface. This intensity fingerprint reveals the extent of the pathology, which leads to filtering the point cloud by those intensity values to extract and calculate the surface defect. On this basis, TLS is proven to be useful to record, detect, characterise, and examine specific building surface deficiencies and carry out the conservation status analysis of the assets surveyed. The case studies in this chapter are heritage buildings with clear surface pathologies. However, given the relationship between the building surface deficiencies and the point cloud data intensities, this research can also be applied to detect anomalies in modern buildings and constructions. D. Antón (B) Research Group ‘TEP970: Innovación Tecnológica, Sistemas de Modelado 3D y Diagnosis Energética en Patrimonio y Edificación’, Departamento de Expresión Gráfica e Ingeniería en la Edificación, Escuela Técnica Superior de Ingeniería de Edificación, Universidad de Sevilla, 4A Reina Mercedes Avenue, 41012 Seville, Spain e-mail: [email protected] Department of Product Design, School of Architecture, Design and the Built Environment, Nottingham Trent University, 50 Shakespeare Street, Nottingham NG1 4FQ, UK M. J. Carretero-Ayuso Departamento de Arquitectura, Universidad de Alcalá, 8 Santa Úrsula Street, 28801 Alcalá de Henares, Spain J. Moyano-Campos ·J. E. Nieto-Julián Departamento de Expresión Gráfica e Ingeniería en la Edificación, Escuela Técnica Superior de Ingeniería de Edificación, Universidad de Sevilla, 4A Reina Mercedes Avenue, 41012 Seville, Spain © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 D. Bienvenido-Huertas and J. Moyano-Campos (eds.), New Technologies in Building and Construction, Lecture Notes in Civil Engineering 258, https://doi.org/10.1007/978-981-19-1894-0_12 207 208 D. Antón et al. Keywords Remote sensing ·Terrestrial laser scanning ·Intensity fingerprint · Point cloud ·Visualisation ·Surface deficiencies ·Conservation status analysis · Pathology ·Heritage building 1 Introduction Building pathologies are diverse in terms of their originating action, their type and extent,andthe characteristicsof thematerialsused.Thesedefectsoranomalies,being either of a physical–mechanical, chemical or biological nature or a combination of them [23], can occur from the moment the buildings are erected. Not only do these problems arise from the interaction of individual components but also due to the interaction of the building with its occupiers and its environment, as well as because of the exposure time [24]. Likewise, other factors such as malpractice in building design, construction, and maintenance are also the cause of building pathologies [5]. As a consequence of their age, historic buildings and constructions are more likely to present surface deficiencies, but modern buildings are not exempt from this. In the scientific community, the use of massive data capture technologies [21] has increased in the latest decades. These systems include terrestrial laser scanning (TLS), the structure-from-motion (SfM) photogrammetric technique, mechanical scanning sonar (MSS), infrared thermography (IRT), among others. Being also intended for diverse purposes, remote sensing is used to and capable of recording the deficiencies in a short time and with much more significant detail than using traditional methods. What is more, irregularities inherent to building deficiencies could only be recorded by these means, which in turn enables the as-built modelling and accurate analysis of the geometrical alterations of heritage assets [2,3,10]. In this sense, Dias et al. [6] carried out a thorough analysis of the scientific literature on the use of advanced digitisation techniques for the detection and diagnosis of building anomalies. The authors concluded that, although these technologies help the visual assessment building façades, there are still great efforts to make in terms of accuracy and reliability in the analysis. Previous research publications in the field have addressed the implementation of TLS and unsupervised classification methods into the mapping of defects of historic buildings via the study of point cloud intensities [4]. Having carried out the classification of defects, Kim et al. [16] presented an automatic diagnosis approach based on TLS for the localisation and quantification of concrete spalling on structures. Masiero and Costantino [22] implemented statistical filtering and machine learning to detect and calculate millimetric damages on brick surfaces of façades from TLS point cloud data. Instead of using cloud intensities, the researchers focused on the computation of distances between points of the brick damaged area and a reference plane on the façade surface. Also working on TLS data via machine learning, Valero et al. [28] processed geometric and SfM colour data to automatically identify and classify material loss defects and discolouration on ashlar structures. Suchocki et al. [26] proposed a method to identify and validate construction surface defects by Laser Scanning Intensity Fingerprint: 3D Visualisation and … 209 using TLS radiometric data (laser intensity). The method was also useful to downsize point clouds via eliminating redundant information, which has an impact on both the computational efforts in processing large datasets and the conservation of spatial data of interest. However, subsampling the point clouds to such an extent (5%) implies that the total number of points affected by the pathology cannot be measured as in this research, nor the full pathology can be mapped. More related to the cataloguing of façade materials than to the analysis of building surface deficiencies, Lezzerini et al. [20] used TLS to record the masonry of the Italian St. Nicholas Church in Pisa. Next, using a Geographical Information System (GIS), these authors mapped the façade components with a view to providing essential information as regards their materials and construction techniques. On the other hand, Antón and Amaro-Mellado [1] highlighted the importance of the digital level (digital number, DN) of the images in the analysis of data. Instead of TLS, they focused on a different remote sensing data source, IRT, also called the infrared thermovision technique. Utilising a GIS and 3D point cloud data processing software, these researchers made use of the DN of the infrared images to produce 3D thermal data to represent the thermal behaviour of buildings and bodies. In view of all the above, considering the capability of TLS to produce intensity data, the research objective is to explore the characterisation of building surface deficiencies through point cloud intensities. In other words, the aim is to analyse the laser scanning intensity fingerprint of building pathologies on surfaces for their detection and quantification. This chapter is structured as follows: Firstly, the case studies are presented. Next, the Methodology section explains the approach, equipment, technologies, and processes considered. Following the discussion of the results achieved, the conclusions and future work take place. 2 Case Studies This section briefly introduces the case studies through which the research objective will be fulfilled. In this chapter, the focus will be on sectors of building components or constructions with visible surface deficiencies caused by different agents. The description of each case study includes concise and diverse details of their history, features, surroundings, and current conservation status. 2.1 St. Andrew’s with Castle Gate United Reformed Church, Nottingham The formerly Presbyterian and now congregation called St. Andrew’s with Castle Gate United Reformed Church [25], known as St. Andrew’s with Castle Gate URC or 210 D. Antón et al. Fig. 1 Colour change in northwest faces of buttresses commonly and simply as St. Andrew’s Church, is registered by the Charity Commission, with the attached former Sunday school, with number 1129944. This institution is based in Nottingham, United Kingdom, in a Grade II listed heritage building (List Entry Number: 1270832), dated 1869 [11]. This church in the Gothic Revival style is surrounded by a wall with railings, also Grade II listed (List Entry Number: 1247549) by Historic England [12]. The church is located opposite the western side of the main buildings of the City Campus of Nottingham Trent University. In relation to the conservation status of the case study, as shown in Fig. 1, both the church’s straight buttresses and the boundary wall evidence a change in the colour of the masonry work in comparison with other areas without this pathology or presenting minor surface deficiencies. In this way, the vertical plane of the buttress and the wall sector on the right (except for the window area) were considered for this research as the affected area and the reference zone for comparison, respectively. 2.2 Bromley House, Beeston, Nottingham There are numerous buildings, elements, and parks and gardens, or parts of them, that have been listed by Historic England throughout the country since 1882. A search in the city of Nottingham alone returns 822 results for heritage assets [13]. It is the case of a residential building in the Victorian style called Bromley House, in Laser Scanning Intensity Fingerprint: 3D Visualisation and … 211 Fig. 2 Bromley House. Efflorescence on the lower part of the western wall. Source Google Street View [8] Beeston, that was the former Board School that can be seen in Fig. 2. This building is located opposite the eastern side of St. John the Baptist Parish Church, also known as Beeston Parish Church, a mid-nineteenth-century Anglican church. This heritage building, listedasGradeII(List EntryNumber:1263823)[14],has undergoneseveral restorations and extensions in the course of its history, such as those of the tower, the bells, the roofs, and the walls [27]. Regarding Bromley House, the building area incorporates the Village Cross, the shaft of the fourteenth-century cross of the village centre, a Grade II listed (List Entry Number: 1263870) heritage asset [15] that was re-erected there in 1929. The part of Bromley House addressed in this research is the lower section of the ground floor western façade, which is made of traditional red brick. As a result of capillarity humidities, the façade shows efflorescence in diverse degrees, ranging from low-density areas to those with intense white stains, together with a slight loss of mass in some bricks. 3 Methodology The recording and processing of remote sensing data of building surface deficiencies are described in this section. The materials and methods involved consist of the approach adopted in this research, together with the equipment, technologies, and processes implemented. Thus, this section is subdivided into two different stages, Data collection, and Data processing. 212 D. Antón et al. 3.1 Data Collection In this way, it is essential to describe the arrangement in the 3D surveys and the setting of parameters. In this subsection, the TLS equipment used to capture the geometry of the two case studies consisted of a tripod-mounted Leica ScanStation P20 [17] 3D laser scanner. The scanner positions (known as stations) and parameters set for each one will be indicated separately. •St. Andrew’s Church The resolution of the laser scanner was set to 3.1 mm at 10 m distance, and the 3D survey was configured as scan only (no images were taken). The number of points (horizontal ×vertical) for that resolution was 20,107 ×8099. The scan quality was set to 3 out of 4. The 3D survey consisted of two scans. With a view to easing the overlap of scan data, the stations were placed 1.5 m apart at 5.5 and 6 m distance from the buttress vertical plane. Due to the proximity of the stations, no targets were used in the laser scanning. The scanning time was 27 min. The time to set the second station is not considered. •Bromley House The resolution of the laser scanner was set to 3.1 mm at 10 m distance, and the 3D survey was configured as scan only (no images were taken). The number of points (horizontal ×vertical) for that resolution was 20,107 ×8099. The scan quality was set to 3 out of 4. The 3D survey consisted of a single scan (no registration process was needed). The scanning time was 13 min and 30 s. 3.2 Data Processing Once the geometric data were recorded, Leica Cyclone 9.4 software [18] was used to produce the case studies’ point clouds. To do this, the scan data were imported into the application. As seen in Sect. 3.1, no images were recorded during the 3D surveys, which means that no textures had to be mapped onto the point clouds. This research analyses the laser beam intensities instead of RGB (red–green–blue) coloured textures. In addition, only the St.Andrew’s Church case study required scan registration (alignment of scans into the same coordinate system) since it consisted of two stations. This was carried out automatically in the software by creating one cloud constraint, but the optimisation of the alignment was also conducted to minimise the group error. Besides, it should be noted that the TLS data were recorded and produced in a local coordinate system for each case study. In other words, no geo-referencing process was performed. Given that the analysis carried out focuses on the geometry and the laser beam intensity data on surfaces, the geo-referencing of the point clouds would not make any difference to the results of this research. Laser Scanning Intensity Fingerprint: 3D Visualisation and … 213 After the point cloud data were produced and stored in .e57 file format, these datasets were manually segmented to select the desired sector to be studied in each case. Using the open-source CloudCompare software [7], this was achieved by conducting manual polygon fencing in the elevation view (XZ plane orthogonal view). Thus, the undesired areas were eliminated to extract the segmented point clouds: the pathology sector (1) and the reference zone (2). In addition, the TLS data were subsampled at 20 mm resolution to try to homogenise the unordered datasets, thus reducing their information entropy [9]. The methodology of this research also involved various processes such as: enquiring the TLS data intensities; producing histograms of the distribution of points among intensity intervals; gathering statistical data from those histograms; filtering the segmented point clouds by laser beam intensity values to isolate the affected areas from the rest of the sectors; likewise, presenting the point cloud data intensities in different visualisation modes with a view to highlighting the affected areas. The processes above were also carried out in CloudCompare software. 4 Results and Discussion This section is structured into two different subsections. Firstly, the outcomes deriving from the research objectives are presented and discussed. This involves showing the digitised areas of the case studies, carrying out the analysis of the TLS data and discussing the findings achieved. Secondly, the research limitations are described and included in a separate subsection. Prior to addressing this section, it is worth providing the main specifications of the computer used to process the data. It was a gaming laptop of the latest generation with an octa-core processor with hyper-threading at 2.30 GHz and a maximum turbo frequency of 4.60 GHz with 24 MB cache, 32 GB RAM DDR4 @ 3200 MHz, a 192- bit graphics card with 3840 cores @ 1425 MHz and 6144 MB GDDR6 dedicated memory @ 14 Gbps, and an NVMe PCIe Gen3x4 SSD (solid-state drive) with 1 TB storage capacity. 4.1 Laser Beam Intensity Fingerprint The research objective is to analyse the intensities of point clouds so that building surface deficiencies can be characterised. Given that the reflection of the laser beam varies depending on the surface colour, material, texture, and humidity content, the buttress of St. Andrew’s Church and the lower wall sector of Bromley House are suitable case studies to reveal the laser beam intensity fingerprint since they present surface deficiencies. 214 D. Antón et al. 4.1.1 St. Andrew’s with Castle Gate URC As mentioned before, the change in the buttress colour can be recorded using TLS and next producing the point cloud data enhanced with intensities. Firstly, it is worth specifying the registration diagnostics of the two scans of the St. Andrew’s Church case study: •Mean absolute error for enabled constraints: 0.001 m •Overlap point count: 100,266 points •Overlap error statistics: – RMS (Root Mean Square) error: 0.0100678 m – Average: 0.00559179 m – Minimum: 1.04814e−06 m – Maximum: 0.0908525 m. Based on both the above data and the validity of the alignment, the next step was to use CloudCompare software to enquire the intensity values from the segmented point clouds: (1) pathology sector and (2) reference zone. Fig. 3 Segmented pathology buttress sector enhanced with intensities: ablue–green–yellow–red; bgreyscale (monochrome); chigh contrast; dmost severely affected part, in monochrome; ethe rest of the sector, in monochrome Laser Scanning Intensity Fingerprint: 3D Visualisation and … 215 Fig. 4 Greyscale histogram of points and intensities of the pathology sector (1) Figure 3displays the segmented area of the pathology (1) dataset enhanced with intensities—in different visualisation modes—where the most affected parts are shown in blue, dark grey, and green colours. Theanalysisofthepoint cloudintensitieswascarriedout considering3500classes or intervals. This amount was chosen so that the width of each interval is reduced in order to increase the accuracy of the graph. Figure 4shows the histogram of point cloud data intensities for the pathology sector (1). The analysis yielded the following data: •556,171 values (points displaying intensity); •Mean: 158.91 points (for each intensity value); •Standard deviation: 213.56 points; •Point of Interest (PoI): 0.25 intensity value. The intensity unit is a coefficient, from 0 to 1 (0 to 100%). Throughout the histogram in Fig. 4, the point arrangement is not symmetrical; i.e. at lower intensities, the point count rate increases more significantly than it decreases at higher intensities. In view of Fig. 3, that is a consequence of the dark areas on the buttress surface. In this way, by mirroring the highest intensities, the ending intensity value of the pathology was established at 0.25 (PoI) for being the symmetry breaking point. Filtering the TLS point cloud data by the intensity PoI, the dataset is split. The remaining 123,498 points with an intensity value lower than 0.25 correspond to the surface deficiency, which constitutes 22.2% of the total number of points in the pathology sector (1). On the other hand, Fig. 5displays the segmented area of the reference (2) dataset enhanced with intensities in the monochrome visualisation mode. The reference sector was rectified for comparison by excluding minor pathological zones on it. 222 D. Antón et al. 4. Armesto-González J, Riveiro-Rodríguez B, González-Aguilera D, Rivas-Brea MT (2010) Terrestrial laser scanning intensity data applied to damage detection for historical buildings. J Archaeol Sci 37:3037–3047. https://doi.org/10.1016/j.jas.2010.06.031 5. DelgadoJMPQ(2021)Casestudiesofbuildingrehabilitationanddesign.SpringerInternational Publishing, Cham 6. Dias I, Flores-Colen I, Silva A (2021) Critical analysis about emerging technologies for building’s façade inspection. Buildings 11:53. https://doi.org/10.3390/buildings11020053 7. Girardeau-Montaut D (2016) CloudCompare: 3D point cloud and mesh processing software. Open Source Project 8. Google Inc. (2021) Street view. In: What is str. view? https://www.google.com/intl/eng/street view/. Accessed 22 Dec 2021 9. He K, Jia MT, Mei-Fang C (2020) Entropy evaluation method for sandstone uranium reservoir characteristics based on convex hull search. IEEE Access 8:46307–46323. https://doi.org/10. 1109/ACCESS.2020.2978153 10. Hichri N, Stefani, De Luca, Veron P (2013) Review of the «As-Built Bim» approaches. Int Arch Photogramm Remote Sens Spat Inf Sci ISPRS Arch XL-5/W1:107–112. https://doi.org/ 10.5194/isprsarchives-XL-5-W1-107-2013 11. Historic England (1995) Church of St. Andrew’s and attached former Sunday school, Non Civil Parish—1270832. In: Listing. https://historicengland.org.uk/listing/the-list/list-entry/127 0832. Accessed 6 Aug 2021 12. Historic England (1995) Boundary wall, railings and gates to Church of St Andrew, Non Civil Parish—1247549. In: Listing. https://historicengland.org.uk/listing/the-list/list-entry/124 7549. Accessed 21 Dec 2021 13. Historic England (2021) Search the list—find listed buildings, monuments, battlefields and more. In: Listing. https://historicengland.org.uk/listing/the-list/. Accessed 22 Dec 2021 14. Historic England (1987) Church of St John the Baptist, non civil Parish—1263823. In: Listing. https://historicengland.org.uk/listing/the-list/list-entry/1263823. Accessed 7 Dec 2021 15. Historic England (1987) Village cross, non civil Parish—1263870. In: Listing. https://histor icengland.org.uk/listing/the-list/list-entry/1263870. Accessed 21 Dec 2021 16. Kim M, Sohn H, Chang C-C (2015) Localization and quantification of concrete spalling defects using terrestrial laser scanning. J Comput Civ Eng 29:04014086. https://doi.org/10.1061/(ASC E)CP.1943-5487.0000415 17. Leica Geosystems (2012) Leica ScanStation P20—industry’s best performing ultra-high speed scanner. In: Scanners. http://w3.leica-geosystems.com/downloads123/hds/hds/ScanStation_ P20/brochures-datasheet/Leica_ScanStation_P20_DAT_us.pdf. Accessed 15 June 2021 18. Leica Geosystems (2019) Leica cyclone—3D point cloud processing software 19. Lerma C, Mas Á, Gil E et al (2014) Pathology of building materials in historic buildings. Relationship between laboratory testing and infrared thermography. Mater Constr 64:e009. https://doi.org/10.3989/mc.2013.06612 20. Lezzerini M, Antonelli F, Columbu S et al (2016) Cultural heritage documentation and conservation: three-dimensional (3D) laser scanning and geographical information system (GIS) techniques for thematic mapping of facade stonework of St. Nicholas Church (Pisa, Italy). Int J Archit Heritage 10:9–19. https://doi.org/10.1080/15583058.2014.924605 21. Mallafrè Balsells C, López Besora JM, Costa Jover A, Coll Pla S (2021) Register of dry stone domes. simplified method for point clouds. Nexus Netw J 23:493–506. https://doi.org/10.1007/ S00004-020-00533-W/FIGURES/8 22. Masiero A, Costantino D (2019) TLS for detecting small damages on a building façade. Int Arch Photogramm Remote Sens Spat Inf Sci XLII-2/W11:831–836. https://doi.org/10.5194/ isprs-archives-XLII-2-W11-831-2019 23. Sitzia F, Lisci C, Mirão J (2021) Building pathology and environment: weathering and decay of stone construction materials subjected to a Csa mediterranean climate laboratory simulation. Constr Build Mater 300:124311. https://doi.org/10.1016/j.conbuildmat.2021.124311 24. Smith M, Gorse C (2021) Building surveyor’s pocket book, 1st edn. Routledge, Abingdon, Oxon; New York, NY Laser Scanning Intensity Fingerprint: 3D Visualisation and … 223 25. St. Andrew’s with Castle Gate United Reformed Church (2021) St. Andrew’s with castle gate United Reformed Church. In: St. Andrew’s with castle gate United Reform. Church—living God’s love through faith action. https://standrewswithcastlegate.org.uk/. Accessed 6 Aug 2021 26. Suchocki C, Błaszczak-B˛ak W, Janicka J, Dumalski A (2021) Detection of defects in building walls using modified OptD method for down-sampling of point clouds. Build Res Inf 49:197– 215. https://doi.org/10.1080/09613218.2020.1729687 27. The Southwell and Nottingham Church History Project (2021) Beeston St. John the Baptist. In: Southwell Nottingham Church historty project. https://southwellchurches.nottingham.ac. uk/beeston/hintro.php. Accessed 7 Dec 2021 28. Valero E, Forster A, Bosché F et al (2019) Automated defect detection and classification in ashlar masonry walls using machine learning. Autom Constr 106:102846. https://doi.org/10. 1016/j.autcon.2019.102846