UAV-Based Structural Damage Mapping: A Review


Structural disaster damage detection and characterization is one of the oldest remote sensing challenges, and the utility of virtually every type of active and passive sensor deployed on various air- and spaceborne platforms has been assessed. The proliferation and growing sophistication of unmanned aerial vehicles (UAVs) in recent years has opened up many new opportunities for damage mapping, due to the high spatial resolution, the resulting stereo images and derivatives, and the flexibility of the platform. This study provides a comprehensive review of how UAV-based damage mapping has evolved from providing simple descriptive overviews of a disaster science, to more sophisticated texture and segmentation-based approaches, and finally to studies using advanced deep learning approaches, as well as multi-temporal and multi-perspective imagery to provide comprehensive damage descriptions. The paper further reviews studies on the utility of the developed mapping strategies and image processing pipelines for first responders, focusing especially on outcomes of two recent European research projects, RECONASS (Reconstruction and Recovery Planning: Rapid and Continuously Updated Construction Damage, and Related Needs Assessment) and INACHUS (Technological and Methodological Solutions for Integrated Wide Area Situation Awareness and Survivor Localization to Support Search and Rescue Teams). Finally, recent and emerging developments are reviewed, such as recent improvements in machine learning, increasing mapping autonomy, damage mapping in interior, GPS-denied environments, the utility of UAVs for infrastructure mapping and maintenance, as well as the emergence of UAVs with robotic abilities.

Practical performance-based design of friction pendulum bearings for a seismically isolated steel top story spanning two RC towers

A case study of performance-based design is presented for a seismically isolated steel structure that rests on top of two adjacent high-rise reinforced concrete towers, the latter separated by means of an expansion joint. The isolation system comprises Friction Pendulum Bearings (FPBs) that are designed to accommodate two salient characteristics of the system. First, the isolated top floor is subjected to narrow-band floor acceleration histories as the ground motion excitation is filtered by the dynamic response of the supporting towers. Second, the displacement demands imposed to the FPBs are affected by the in-phase or out-of-phase movement of the supporting structures, with the latter case potentially giving rise to higher displacement capacity requirements for the bearings. In a search for a solution beyond conventional design norms, the probability of bearing failure associated with a wide range of FPB displacement capacities was determined via an explicitly risk-consistent performance-based seismic design. Overall, the case-specific design approach is shown to be able to meet any desired performance objective, consistently determining the final compromise between safety, cost-efficiency and practicability.

Combined Convolutional Neural Networks and Fuzzy Spectral Clustering for Real Time Crack Detection in Tunnels


A computer vision module is proposed for crack detection in tunnels, a challenging process due to the low visibility, the curvature from, and the structures of the cracks which, though being very narrow in width, they are very deep. Our system is embedded on a robot which surveys tunnels in real-time as it is moving in the infrastructure. Initially, a Convolutional Neural Network is employed to detect the cracks which, however, yields only approximate regions due to the great complexity of the scene. Then, a combined fuzzy spectral clustering is then introduced to refine the detected crack regions exploiting spatial and orientation coherency. The algorithms have been tested in real-life tunnels in Egnatia Highway. Our scheme yields high detection accuracy than existing methods and the capacity of the robot to touch the crack to allow in-situ measurements within a precision of 2-3cm in a tunnel of 7m height.

A Seismic Design Procedure for Different Performance Objectives for Post-Tensioned Walls


A method is presented for the design of unbonded post-tensioned concrete walls for seismic loading to satisfy multiple performance objectives. It takes advantage of the fact that the initial stiffness of the wall is nearly independent of the amount of post-tensioning reinforcement, and thus the fundamental period of the building can be considered to be a stable parameter in design of walls of a given cross section, independent of the degree of post-tensioning. The design spectra used to follow the specification provided in Eurocode 8 considering the probability of exceedance of the seismic action. A detailed example is provided.

Aumento de la Resiliencia de las Carreteras Mediante el Uso Combinado de Tecnología Multisensor y Modelos Climáticos


Le projet PANOPTIS, financé par la Commission européenne dans le cadre du programme H2020, vise à accroître la résilience (capacité d’adaptation) des routes aux conditions météorologiques défavorables, telles que les phénomènes météorologiques extrêmes ou les inondations, et à d’autres événements à risque comme les tremblements de terre ou les glissements de terrain. L’objectif principal du projet est de combiner des scénarios de changement climatique régionalisés (appliqués aux infrastructures), avec des outils de simulation structurels et géotechniques et des données réelles tirées directement des infrastructures routières (ponts, pentes, routes) par un réseau multi-capteurs comprenant des capteurs au sol, des drones et des satellites, afin de fournir aux gestionnaires d’infrastructures de transport un outil de contrôle intégré, capable d’améliorer la gestion des infrastructures dans les phases de planification, de maintenance et d’exploitation. Le projet PANOPTIS a débuté en juin 2018, et pendant la première phase du projet, qui couvre approximativement les deux premières années, toutes les technologies innovantes qui composent l’outil PANOPTIS sont en cours de mise au point. Au cours de la deuxième phase du projet, qui débutera à l’été 2020, ACCIONA Engineering mettra en oeuvre toutes les technologies et méthodologies développées dans la section 2 de l’autoroute A-2, qui est longue de 77,5 km et qui traverse la province de Guadalajara. Il s’agit d’une section de concession d’autoroute de première génération gérée par ACCIONA Concessions et entretenue par ACCIONA Maintenance.

Pixel-Level Corrosion Detection on Metal Constructions by Fusion of Deep Learning Semantic and Contour Segmentation


Corrosion detection on metal constructions is a major challenge in civil engineering for quick, safe and effective inspection. Existing image analysis approaches tend to place bounding boxes around the defected region which is not adequate both for structural analysis and pre-fabrication, an innovative construction concept which reduces maintenance cost, time and improves safety. In this paper, we apply three semantic segmentation-oriented deep learning models (FCN, U-Net and Mask R-CNN) for corrosion detection, which perform better in terms of accuracy and time and require a smaller number of annotated samples compared to other deep models, eg CNN. However, the final images derived are still not sufficiently accurate for structural analysis and pre-fabrication. Thus, we adopt a novel data projection scheme that fuses the results of color segmentation, yielding accurate but over-segmented contours of a region, with a processed area of the deep masks, resulting in high-confidence corroded pixels.

Multi-Label Deep Learning Models for Continuous Monitoring of Road Infrastructures


A multi-class, multi-label deep learning model for the monitoring of road infrastructures is presented in this paper. The employed detection methodology can identify animals, debris, road defects, fire, fog, flooded areas and humans. All these categories are strongly related to the efficient movement of vehicles through a transportation network. Possible detections indicate roadway disruptions of various types. Therefore, they should be detected as fast as possible. Experimental results indicate that the proposed scheme presents high detection results and, thus, can be used in any motorway monitoring process.

Downtime Assessment of Base-isolated Liquid Storage Tanks


Seismic base isolation is examined as a design alternative for supporting industrial facility liquid storage tanks against earthquake loading. A 160,000 m3 liquid storage tank is adopted as a case study, for which two designs are assessed, one with and one without base isolation. Using a nonlinear surrogate model and a set of ground motion records selected using the conditional spectrum approach for the average spectral acceleration intensity measure, Incremental Dynamic Analysis is employed to derive seismic fragility curves. Consequences of damage are evaluated in terms of downtime, considering the characteristics of petrochemical storage tanks, whereby any repair requires a lengthy list of actions dictated by health and safety requirements. The results reveal considerable benefits when base-isolation is employed, by drastically reducing downtime when sufficient displacement capacity is provided in the isolators.

Mixed Probabilistic Seismic Demand Models for Fragility Assessment

A mixture model is presented for combining the results of different models or analysis approaches into a single probabilistic seismic demand model that is suitable for fragility assessment. A structure can be represented using different model types or even levels of resolution for the same type, while it may also be analysed via methods of different complexity, most notably static versus dynamic nonlinear approaches. Combining the results from different models or analysis methods can be beneficial as it allows updating the results of a simpler approach or combining the strengths of two different models. For example, different model types may offer accuracy advantages in complementary response regions. This is the case of distributed-plasticity fiber models that offer higher fidelity for reinforced concrete frames at low (pre-capping) deformations, while lumped-plasticity models are more reliable for larger (postcapping) deformations closer to collapse. Through the combination of the results of both models we can potentially better capture the performance of the frame at all levels of seismic intensity. By employing a minimal 5 parameter power-law-based model we offer viable options for forming mixed probabilistic seismic demand models that can combine both different models and different analysis methods into a single output suitable for fragility assessment.

Decision Support for Road Infrastructure Resilience: The PANOPTIS Perspective


The PANOPTIS consortium aims to leverage existing tools and services as well as remote sensing technologies to deliver an integrated platform that can address road infrastructure (RI) multi-hazard resilience. The scope of the project incorporates RI structural components (bridges, overpasses, interchanges, tunnels, slopes, retaining walls, pavements, and surface water drains), non-structural components (tunnel ventilation systems, traffic cameras and signposts), as well as interconnected non-RI components, such as power transmission lines and telecommunication towers. Both detailed and surrogate structural models will be developed for RI and non-RI components, quantifying and incorporating the epistemic uncertainty due to the detailed models’ reduction to surrogacy to allow a rapid high-resolution assessment of vulnerability, whereby loss, functionality and downtime become directly tied to rehabilitation/emergency action planning. The focus is on the development of a rapid-response decision-support tool that will employ measured data immediately after any seismic event to issue inspection prioritization protocols, facilitate the rapid assessment of the state of the RI, and help increase its resilience to catastrophic events.

Prescriptive Approaches in Performance-based Design? A Case-study on Base Isolation

The collapse performance of code-designed base-isolated structures has recently received considerable criticism, having been found to be deficient vis-à-vis conventional buildings in several situations. As a remedy, prescriptive minima with a tenuous probabilistic justification have been recommended in the literature for the bearing deformation capacity. These are independent of structure or site characteristics, yet they are already finding use in design. We put this concept to the test by means of a case study of a seismically isolated steel structure that rests on the roof of two adjacent high-rise reinforced concrete towers. To seismically isolate the steel structure, Friction Pendulum Bearings (FPBs) are used, and their displacement capacity is determined to comply with a performance objective of 1% probability of collapse in 50 years. The case study possesses two salient features that distinguish it from pertinent past investigations. The first is that the isolated steel structure rests on top of two others and consequently it is subjected to narrow-band roof acceleration time histories, shaped by the filtering of the ground motion excitation through the supporting buildings. The second is that the two supporting towers have different modal characteristics, thus displacement demands imposed to the FPBs are mainly affected by their in-phase or out-of-phase movement. Overall, a case-specific true performance-based design is shown to achieve the desired safety while requiring 1.5 times lower displacement capacities for the bearings, when compared to prescriptive “performance-based” approaches.

Issues in Harmonization of Seismic Performance via Risk Targeted Spectra


Current seismic design code provisions are mainly based on checking structural performance at a single seismic intensity associated with a pre-defined return period. For instance, in EN1998, a ground motion with 10% probability of exceedance in 50 years is used for design. This design procedure, with the inclusion of partial safety factors, is assumed to provide sufficient safety margin against earthquakes for newly designed buildings. Nevertheless, it does not specifically determine the expected seismic risk related to any performance level or limit state. Therefore, it may result in non-uniform risk for buildings located in different sites within a region (or country), even for places with identical design intensities. Instead, ASCE 7-10 incorporates Risk Targeted design maps that suggest the application of suitable spectra adjustment factors, in order to ensure a reasonably low uniform collapse risk. Making use of simplified single degree of freedom structures defined in several configurations of period and ductility, our aim is to test the effectiveness of the adjustment factors computed under different assumptions. It is shown that, although matching is not practically possible, harmonization remains a viable target, offering insights for possible future adoption of Risk Targeted Spectra in forthcoming seismic codes.

Improving Resilience of Transport Instrastructure to Climate Change and other natural and Manmande events based on the combined use of Terrestrial and Airbone Sensors and Advanced Modelling Tools


The project PANOPTIS, funded by the European Commission under the H2020 Programme, aims at increasing the resilience of the transport infrastructures (focusing on roads) and ensuring reliable network availability under unfavourable conditions, such as extreme weather, landslides, and earthquakes. The main target is to combine downscaled climate change scenarios (applied to road infrastructures) with structural and geotechnical simulation tools and with actual data from a multi-sensor network (terrestrial and airborne-based), so as to provide the operators with an integrated tool able to support more effective management of their infrastructures at planning, maintenance and operation level. During the first stage of the project, the consortium will develop advanced technologies to monitor and control transport infrastructures, such as a Geotechnical and Structural Simulation Tool (SGSA) to predict structural and geotechnical risks in road infrastructures; drone-technologies applied to road upkeep and incident management; improved computer vision and machine learning techniques for damage diagnosis of infrastructure, and early warning systems to help operators identify and communicate emerging systemic risks. At the same time, experts in climate modelling, will analyse the possible short and long term effects of climate change on transport infrastructure (e.g. flooding, heavier snows). All the information from the different sensors, models and applications will be integrated and processed through a unique Resilience Assessment Platform that will support operators in the introduction of adaptation and mitigation strategies based on multi-risk scenarios. During the second stage of the project, ACCIONA Engineering will implement the developed technologies and methodologies in a section of the Spanish A-2 motorway, in the province of Guadalajara. PANOPTIS integrated Platform will help optimize the management and maintenance of the Ministry of Public Works’ concession for a 77.5-km section, all in collaboration with ACCIONA Infrastructure Maintenance (AMISA) and ACCIONA Concessions. In parallel, PANOPTIS platform will also be implemented in a section of 62 Km of a Greek motorway, renowned for its seismic activity. The trials in Greece hosted by the operator Egnatia Odos will integrate the motorway that serves the Airport of Thessaloniki. So the scenario will integrate a modal transfer segment.

Infrastructure Degradation And Post-disaster Damage Detection Using Anomaly Detecting Generative Adversarial Networks


Degradation and damage detection provides essential information to maintenance workers in routine monitoring and to first responders in post-disaster scenarios. Despite advance in Earth Observation (EO), image analysis and deep learning techniques, the quality and quantity of training data for deep learning is still limited. As a result, no robust method has been found yet that can transfer and generalize well over a variety of geographic locations and typologies of damages. Since damages can be seen as anomalies, occurring sparingly over time and space, we propose to use an anomaly detecting Generative Adversarial Network (GAN) to detect damages. The main advantages of using GANs are that only healthy unannotated images are needed, and that a variety of damages, including the never before seen damage, can be detected. In this study we aimed to investigate 1) the ability of anomaly detecting GANs to detect degradation (potholes and cracks) in asphalt road infrastructures using Mobile Mapper imagery and building damage (collapsed buildings, rubble piles) using post-disaster aerial imagery, and 2) the sensitivity of this method against various types of pre-processing. Our results show that we can detect damages in urban scenes at satisfying levels but not on asphalt roads. Future work will investigate how to further classify the found damages and how to improve damage detection for asphalt roads.

Post-Disaster Building Damage Detection from Earth Observation Imagery Using Unsupervised and Transferable Anomaly Detecting Generative Adversarial Networks


We present an unsupervised deep learning approach for post-disaster building damage detection that can transfer to different typologies of damage or geographical locations. Previous advances in this direction were limited by insufficient qualitative training data. We propose to use a state-of-the-art Anomaly Detecting Generative Adversarial Network (ADGAN) because it only requires pre-event imagery of buildings in their undamaged state. This approach aids the post-disaster response phase because the model can be developed in the pre-event phase and rapidly deployed in the post-event phase. We used the xBD dataset, containing pre- and post- event satellite imagery of several disaster-types, and a custom made Unmanned Aerial Vehicle (UAV) dataset, containing post-earthquake imagery. Results showed that models trained on UAV-imagery were capable of detecting earthquake-induced damage. The best performing model for European locations obtained a recall, precision and F1-score of 0.59, 0.97 and 0.74, respectively. Models trained on satellite imagery were capable of detecting damage on the condition that the training dataset was void of vegetation and shadows. In this manner, the best performing model for (wild)fire events yielded a recall, precision and F1-score of 0.78, 0.99 and 0.87, respectively. Compared to other supervised and/or multi-epoch approaches, our results are encouraging. Moreover, in addition to image classifications, we show how contextual information can be used to create detailed damage maps without the need of a dedicated multi-task deep learning framework. Finally, we formulate practical guidelines to apply this single-epoch and unsupervised method to real-world applications.

Project Interactions

Latest News

Contact us



This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No  769129