Recent years have been fruitful in the development of computer vision methods for a wide variety of applications. Despite the successful results achieved in the segmentation of cracks on concrete surfaces, poor results are still persisting during onsite application, mainly due to noise caused by biological colonization, which is present in most of historical heritage buildings. The authors have been working on this problematic and developed the SC-Crack method previously, however it still relies on the cumbersome task of acquiring sets of 17 channels images to compose an hyperspectral cube and still requires case-wise hyperparameter optimization. Consequently, it is important to define which spectral information mostly defines the success of the method, enabling to optimize both, the acquisition procedure and model processing. Following, a study aiming at the selection of the more informative channels was carried and the hyperparameter-free model is evaluated.In this scope, images of concrete specimens were acquired sequentially to compose a 17 channel hyperspectral image cube. These were sere compute allowing to define the most informative channels sets that are processed using the SC-Crack+ method, presented in this work. The reduced image cubes of cracking on clean concrete surfaces and on surfaces with biological colonization were processed and analyzed. Relevant and improved results were achieved for crack segmentation, following this SC-Crack+ model. This enables the possibility of mounting cameras with sensors and lenses particularly adapted for prone acquisition targeting only the most relevant hyperspectral information for crack segmentation and still using traditional feature engineering image processing methods.

Computer vision enables a much more efficient monitoring system of structural behavior, compared to traditional methods. This derives from the fact that it allows the assessment of displacements in a vast number of points that can be processed to the estimate deformations, accelerations, and other key parameters at relevant cross-sections. This paper proposes a computer vision-based methodology, specifically designed for monitoring seismic tests conducted on a shaking table with a reduced-scale model, using a single camera approach. This innovative methodology uses artificial targets with predefined color and geometry to optimize their detection and tracking. These targets are positioned at key points of the reduced model-"Moving Targets"-and at reference points of the seismic table-"Control Targets." Videos are recorded during the seismic tests and the coordinates (in pixels) of the targets' centers are automatically detected and captured. Then, transformations are applied to each frame to compute the targets coordinates in millimeters and based on the differences between frames, to calculate the displacements of each target. The methodology was first calibrated with dynamic tests performed on a reduced-scale model (1:33) of a prefabricated concrete shell, printed in acrylonitrile butadiene styrene (ABS), conducted on an educational shaking table. Afterwards, the methodology was validated with seismic tests performed on a 1:3 reduced model of a prefabricated concrete shell, produced according to the similitude theory to best reproduce the behavior of the corresponding prototype. It was demonstrated the ability of the methodology to monitoring seismic tests, enabling continuous tracking of the targets placed on a structure with complex geometry.

Maintenance decisions often extend beyond physical deterioration, with stakeholders' experiences playing crucial roles. Traditionally, building assessments rely on visual inspections, but manual methods have limitations (e.g., time and cost expensive, relying on subjective assessment). “Feeling-BIM” proposes a new methodology whose concept integrates automated facade inspection with residents' sentiments. Incorporating residents' feelings acknowledges the prevalence of subjective criteria in maintenance decisions, with a particular focus on facade stains that influence interventions and building aesthetics. Three buildings with the same archetype and in the same street, in different degradation conditions, are analysed as case study. A sentiment analysis is performed, to evaluate the ability of residents to clearly distinguish between facades in the best condition and those identified by automated inspection as having a higher presence of stains. “Feeling-BIM” empowers decision-makers with a more comprehensive perspective on facades’ condition, ensuring that decisions are well-informed and balanced, considering both objective and subjective factors.

Understanding the mechanisms of pipeline failures is crucial for identifying vulnerabilities in gas transmission pipelines and planning strategies to enhance the reliability and resilience of energy supply chains. Existing studies and the American Society of Mechanical Engineers’ (ASME) Code for Pressure Piping primarily focus on corrosion, recommending inspections every 10 years to prevent incidents due to this time-dependent threat. However, these guidelines do not provide comprehensive regulation on the likelihood of incidents due to other causes, especially non-time-dependent events (i.e. do not provide any indication of the inspection frequency or the most likely time for an incident to occur). This study adopts an innovative approach adopting machine learning, particularly artificial neural networks (ANNs), to analyse historical pipeline failure data from 1970 to 2023. By analysing records from the US Pipeline Hazardous Materials Safety Administration, the model captures the complexity of various degradation phenomena, predicting failure years and hazard frequencies beyond corrosion. This innovative approach allows adopting more informed preventive measures and response strategies, offering deep insights into incident causes, consequences, and patterns. The results provide practical insights for maintenance planning, offering an estimation of periods when a pipeline may be more susceptible to incidents based on various factors. However, since all models inherently present uncertainties, both in the data and the modelling process, these estimates should be interpreted as probabilistic assessments. This study provides operators with a strategic framework to prescriptively address potential vulnerabilities, thereby promoting sustained operational integrity and minimising the occurrence of unexpected events throughout the service life of pipelines. By expanding the scope of risk assessment beyond corrosion, this study significantly advances the field of pipeline safety and reliability, setting a new standard for comprehensive incident prevention.

With the rapid development of society and the economy, significant infrastructure, such as roads, buildings, high-speed railways, and bridges, have been built all over the world [...]

Image-based methods have been applied to support structural monitoring, product and material testing, and quality control. Lately, deep learning for compute vision is the trend, requiring large and labelled datasets for training and validation, which is often difficult to obtain. The use of synthetic datasets is often applying for data augmentation in different fields. An architecture based on computer vision was proposed to measure strain during prestressing in CFRP laminates. The contact-free architecture was fed by synthetic image datasets and benchmarked for machine learning and deep learning algorithms. The use of these data for monitoring real applications will contribute towards spreading the new monitoring approach, increasing the quality control of the material and application procedure, as well as structural safety. In this paper, the best architecture was validated during experimental tests, to evaluate the performance in real applications from pre-trained synthetic data. The results demonstrate that the architecture implemented enables estimating intermediate strain values, i.e., within the range of training dataset values, but it does not allow for estimating strain values outside those range. The architecture allowed for estimating the strain in real images with an error ∼0.5%, higher than that obtained with synthetic images. Finally, it was not possible to estimate the strain in real cases from the training performed with the synthetic dataset.

Non-invasive vision-based approaches provide the precision and reliability required for building facade inspection. A method based on automatic image classification to detect and map the materials and anomalies in building facades is presented and compared with a traditional visual inspection and classification.

In the restoration of exposed concrete buildings with architectural or cultural relevance, the chromatic and texture compatibility between the adopted repair mortars and the original concrete surface plays a key factor. In the particular case of smooth white concrete surfaces, this compatibility can be even more challenging. The application of commercial products ensures mechanical aspects, but the aesthetic matching between both materials is hard to achieve. A methodology for restoration of smooth exposed surfaces of white concrete is herein presented. The concrete surface is chromatically characterized by image processing, and then a customized restoration mortar is designed by adding pigment to a reference mortar, defined to fulfil mechanical and durability requirements. The methodology was validated in ‘Pavilhão do Conhecimento’, the first building in exposed white concrete built in Portugal. The chromatic and finishing compatibility was achieved, and the interventions are not perceptible to the human eye. The developments proposed results in a science-based approach that enable the reproduction of smooth white concrete surface finishes and the chromatic reintegration to replicate colour heterogeneity and soften the transition zones.

Structural monitoring plays a crucial role in assessing damage and detecting potential faults. Vision systems provide real-time information about the behaviour and state of conservation of structures, enabling precise and efficient analysis. Computer vision algorithms are employed to analyse images and extract relevant structural information. This information may encompass key points' displacement and deformation within the structure.This paper introduces a methodology for monitoring dynamic tests using computer vision. The approach employs a camera to track artificial targets placed on the structure. The camera captures multiple images during the tests and detects the coordinates of the target centres. Subsequently, displacements are calculated based on the differences in coordinates among the aforementioned targets.To validate the methodology, an experimental study was conducted using dynamic tests on a reduced-scale model of a thin concrete shell tested on a seismic table. Five dynamic tests were carried out, involving varied amplitude and frequency values to test the proposed approach. The results substantiate the efficacy of computer vision in monitoring dynamic tests. In conclusion, the developed methodology facilitates dynamic test monitoring through computer vision by continuously tracking artificial targets throughout the tests and computing their displacements.

The Ocean Swimming Pool (1960–1966) is one of Álvaro Siza’s most internationally recognized works for its exceptional landscape integration while expressing a tectonic shift from regionalist inspiration towards more abstract design and innovative constructive solutions. Recent conservation has enhanced the site’s significance, by preserving the original design principles and extending the building to the north, where the original construction had been left unfinished. Under the Keeping It Modern Grant awarded by the Getty Foundation, inspection and diagnosis were carried out during the building site, providing a complete material assessment of the building. Also, the funding allowed for the localized repair of concrete spalling due to steel corrosion which went beyond the traditional patch repair by applying innovative techniques of chromatic and texture integration between the existing and the new repair mortars.

The pre-wall system can be used in the construction of buildings, with multiple advantages in terms of production time and environmental impact. This type of walls is formed by two thin precast concrete panels, linked with steel trusses to create a hollow precast wall module, which are assembled at construction site, adding reinforcement in critical regions, namely, in connections, and casting concrete into the core. Recently, the durability and sustainability of this wall system were enhanced, using a high-durability concrete only in the cover and for the core was used a concrete that combines low cement dosage with lightweight aggregates. This paper presents a study, with both experimental and numerical components, developed to characterize the behaviour of corner connections between two perpendicular pre-walls. The connections between perpendicular pre-walls are must more complex than traditional connections made in situ because the space to define the reinforcement detailing is very limited, making difficult to assure a proper transmission of the internal stresses between the two walls. Different amounts of reinforcement, concretes mixtures and reinforcement detailing were tested. Nine numerical simulations were also performed to further enrich the results discussion. Cracks pattern, deformation, strength and stiffness were analysed. Conclusions are drawn and guidelines are proposed to design this type of connections between pre-walls.

The new Construction 4.0 paradigm takes advantage of existing technologies. In this scope, the development and application of image-based methods for evaluating and monitoring the state of conservation of buildings has shown significant growth, including support for maintenance plans. Recently, powerful algorithms have been applied to automatically evaluate the state of conservation of buildings using deep learning frameworks, which are utilised as a black-box approach. The large amount of data required for training, the difficulty in generalising, and the lack of parameters to assess the quality of the results often make it difficult for non-experts to evaluate them. For several applications and scenarios, simple and more intuitive image-based approaches can be applied to support building inspections. This paper presents the StainView, which is a fast and reliable method. The method is based on the classification of the mosaic image, computed from a systematic acquisition, and allows one to (i) map stains in facades; (ii) locate critical areas; (iii) identify materials; (iv) characterise colours; and (v) produce detailed and comprehensive maps of results. The method was validated in three identical buildings in Bairro de Alvalade, in Lisbon, Portugal, that present different levels of degradation. The comparison with visual inspection demonstrates that StainView enables the automatic location and mapping of critical areas with high efficiency, proving to be a useful tool for building inspection: differences were of approximately 5% for the facade with the worst and average state of conservation, however, the values deteriorate for the facade under good conditions, reaching the double of percentage. In terms of processing speed, StainView allows a facade mapping that is 8–12 times faster, and this difference tends to grow with the number of evaluated façades.

Maintenance decisions at the end of building components’ service life are frequently driven by subjective motivations that can arise from various sources, including the building owner’s personal preferences, sentimental attachments, aesthetic considerations, and individual/collective preferences or sense of taste. This study supports decision-making regarding maintenance actions by combining objective indicators of building degradation and subjective user perceptions to prioritize areas of focus, determine appropriate maintenance strategies, and allocate resources effectively.

The interface between concrete layers cast at different ages is present both in new construction, e.g., precast beams with cast in place decks, and in rehabilitation, e.g., jacketing of existing beams and columns. The interface strength ensures the element’s monolithic behaviour. This strength is mainly determined by the interface surface roughness. Unlike other material parameters, the interface strength is accessed with deterministic parameters (cohesion and friction) making design dependent on their correct assessment. Within this framework, the research study herein described aims to find non-deterministic parameters of the interface strength and an automatic method for their evaluation. The proposed parameters are derived from best-fit geostatistical models for roughness. These parameters are neither deterministic nor used for roughness and strength evaluation but they describe natural phenomena well. The automatisation is obtained via the deep learning process of the interface texture and strength. Since this process requires huge and not available data, real sources of data (i.e. surfaces) are substituted with the virtual Gaussian models generated with the geostatistical parameters obtained from the real surfaces. A similar process can be applied to the strength data if available in sufficient numbers. For short, the paper presents the application of the geostatistical parameters to the evaluation of both surface texture and strength of concrete to concrete interface, which have not been used before. It is shown that these parameters can be used to ‘store’ surface data in a compact form and subsequently texture or strength virtual reproduction is possible making deep learning and automatisation possible to perform.

The strengthening of concrete structures with laminates of Carbon-Fiber-Reinforced Polymers (CFRP) is a widely adopted technique. retained The application is more effective if pre-stressed CFRP laminates are adopted. The measurement of the strain level during the pre-stress application usually involves laborious and time-consuming applications of instrumentation. Thus, the development of expedited approaches to accurately measure the pre-stressed application in the laminates represents an important contribution to the field. This paper proposes and benchmarks contact-free architecture for measuring the strain level of CFRP laminate based on computer vision. The main objective is to provide a solution that might be economically feasible, automated, easy to use, and accurate. The architecture is fed by digitally deformed synthetic images, generated based on a low-resolution camera. The adopted methods range from traditional machine learning to deep learning. Furthermore, dropout and cross-validation methods for quantifying traditional machine learning algorithms and neural networks are used to efficiently provide uncertainty estimates. ResNet34 deep learning architecture provided the most accurate results, reaching a root mean square error (RMSE) of 0.057‰ for strain prediction. Finally, it is important to highlight that the architecture presented is contact-free, automatic, cost-effective, and measures directly on the laminate surfaces, which allows them to be widely used in the application of pre-stressed laminates.

The development of automatic methods to recognize cracks in surfaces of concrete has been under focus in recent years, firstly through computer vision methods and more recently focusing on convolutional neural networks that are delivering promising results. Challenges are still persisting in crack recognition, namely due to the confusion added by the myriad of elements commonly found on concrete surfaces. The robustness of these methods would deal with these elements if access to correspondingly heterogeneous datasets was possible. Even so, this would be a cumbersome methodology, since training would be needed for each particular case and models would be case dependent. Thus, efforts from the scientific community are focusing on generalizing neural network models to achieve high performance in images from different domains, slightly different from those in which they were effectively trained. The generalization of networks can be achieved by domain adaptation techniques at the training stage. Domain adaptation enables finding a feature space in which features from both domains are invariant, and thus, classes become separable. The work presented here proposes the DA-Crack method, which is a domain adversarial training method, to generalize a neural network for recognizing cracks in images of concrete surfaces. The domain adversarial method uses a convolutional extractor followed by a classifier and a discriminator, and relies on two datasets: a source labeled dataset and a target unlabeled small dataset. The classifier is responsible for the classification of images randomly chosen, while the discriminator is dedicated to uncovering to which dataset each image belongs. Backpropagation from the discriminator reverses the gradient used to update the extractor. This enables fighting the convergence promoted by the updating backpropagated from the classifier, and thus generalizing the extractor enabling it for crack recognition of images from both source and target datasets. Results show that the DA-Crack training method improved accuracy in crack classification of images from the target dataset in 54 percentage points, while accuracy on the source dataset remains unaffected.

Strengthening of reinforced concrete (RC) structures with pre-stressed Carbon Fiber Reinforced Polymer (CFRP) laminates is a well-known application. The development of vision-based approaches for monitoring the strain imposed during the pre-stress application, with the required precision and accuracy, represents an important contribution for the state of the art. A new system, named Strain- Vision, was design and developed tacking into account three main modules: (i) development of a customized high precision strain monitoring CFRP laminates (hpsm-CFRP); (ii) definition of a set-up for image acquisition during pre-stress application; (iii) design of computer vision architecture based on deep learning to measure the strain. The pre-processing of data, to be analysed with an architecture previously training, is herein discussed, aiming to improve the quality and performance of the system without the need for large datasets, usually required in deep learning applications.

Tourists’ perceptions of monuments influence their feelings about the country and the possibility of returning or recommending their visit to other tourists. TripAdvisor is one of the most popular websites for sharing travelling experiences and plays an important role when choosing a travel destination. But what are the factors that can provoke negative feelings in tourists? The maintenance of monuments is essential for their conservation; however, active maintenance can trigger negative feelings in tourists, compromising their connection with the cultural heritage of the country. This study reveals how some maintenance actions can influence tourists’ expectations regarding two relevant architectural monuments in the Iberian Peninsula by applying VADER (Valence Aware Dictionary for sEntiment Reasoning) to 13,000 TripAdvisor reviews written in the last decade and in three languages. Other variables, such as weather conditions and changes in climate, tourists’ country of origin and their style of travel, are evaluated to eliminate the possible mediating effects of these variables. This study reveals that the maintenance status of monuments seems to be the variable with the greatest impact on tourists’ perceptions and on their evaluations on TripAdvisor, propagating negative feelings towards the monument, from which it takes some time to recover.

The conservation of exposed concrete heritage implies repairs with restoration demands, besides mechanical and durability requirements. The main difficulty is to achieve chromatic, texture and finishing compatibility between repair mortars and concrete surfaces. The authors developed a restoration methodology for concrete heritage that comprises the chromatic characterization by image processing and the design of customized restoration mortars, fulfilling mechanical, durability, colour and finishing requirements. In this paper, the methodology is applied to a case-study, proving its ability to achieve the desired compatibility between the original surface and the restoration mortar, thus validating its use in safeguarding concrete heritage.