3D reconstruction of disaster scenes for urban search and rescue (original) (raw)
Related papers
2017
Structural damage assessment (SDA) is a crucial activity during post-disaster response and reconstruction phases. Recent advances in photogrammetry and computer vision permit to obtain very dense 3D point cloud models from overlapping aerial multi-perspective images. Both 3D point clouds and multi-perspective imagery are rich sources of information for damage mapping. Multi-perspective aerial imagery is commonly obtained by Unmanned Aerial Vehicles (UAVs); however, these data can be scarce during crisis situations. An interesting alternative are post-disaster aerial video footages, but low resolution and redundancy of video frames hinders its utility. Exploration of video frames usability for 3D modelling, particularly regarding postdisaster applications, is still lacking. In this research the quality of aerial video-generated 3D models was assessed from geometric/absolute and SDA application perspectives, and was compared with models derived from aerial still imagery for two different study areas. Particularly video blur-motion, resolution and frame redundancy influence on 3D model quality was determined. The analysis demonstrated that in general video data produce more noisy and imprecise 3D point clouds; however, the external and absolute accuracy is still comparable to the one of still imagery. Low resolution video was clearly hampered by sensor proximity to the ground, whereas frame redundancy was the main cause of noise. These quality parameters, however, were also related to higher point density and in most cases better representability of damage-related features. Consequently, it was demonstrated that video data are suitable for the generation of rich damage-related information 3D models.
Rescue Management and Assessment of Structural Damage by Uav in Post-Seismic Emergency
2020
Abstract. The increasing frequency of emergencies urges the need for a detailed and thorough knowledge of the landscape. The first hours after a disaster are not only chaotic and problematic, but also decisive to successfully save lives and reduce damage to the building stock. One of the most important factors in any emergency response is to get an adequate awareness of the real situation, what is only possible after a thorough analysis of all the available information obtained through the Italian protocol Topography Applied to Rescue. To this purpose geomatic tools are perfectly suited to create, manage and dynamically enrich an organized archive of data to have a quick and functional access to information useful for several types of analysis, helping to develop solutions to manage the emergency and improving the success of rescue operations. Moreover, during an emergency like an earthquake, the conventional inspection to assess the damage status of buildings requires special tools...
Uav Strategies Validation and Remote Sensing Data for Damage Assessment in Post-Disaster Scenarios
ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
The recent seismic swarms, occurred in Italy since August 2016, outlined the importance of deepen Geomatics researches for the validation of new strategies aimed at rapid-mapping and documenting differently accessible and complex environments, as in urban contexts and damaged built heritage. In the emergency response, the crucial exploitation of technological advances should obtain and efficiently organize high-scale reliable geospatial data for the early warning, impact, and recovery phases. Fulfilling these issues, among others, the Copernicus EMS, has played by now an important role in immediate and extensive damage reconnaissance, as in the case of Centre Italy. Nevertheless, the use of remote sensing data is still affected by a problem of point-of-view, scale and detectable detail. Nadir images, airborne or satellite, in fact, strongly limited the confidence level of these products. The subjectivity of the operator involvement is still an open issue, both in the first fieldwork assessment, and in the following operational approach of interpretative damage detection and rapid mapping production. To overcome these limits, the introduction of UAV platforms for photogrammetric purposes, has proven to be a sustainable approach in terms of time savings, operators' safety, reliability and accuracy of results: the nadir and oblique integration can provide large multiscale models, with the fundamental information related to the façades conditions. The presented research, conducted within the Central Italy earthquakes events, will focus on potentialities and limits of UAV photogrammetry in the two documented sites: Pescara del Tronto and Accumoli. Here, the aim is not limited to describe a series of strategies for georeferencing, blocks orientation and multitemporal co-registration solutions, but also to validate the implemented pipelines as a workflow that could be integrated in the operative intervention for emergency response in early impact activities. Thus, it would be possible to use this 3D metric products as a reference-data for significative improvements of reliability in typical visual inspection and mapping, flanking the traditional nadir airborne-or satellite-based products. The UAV acquisitions performed in two damaged villages are displayed, in order to underline the implication of the spatial information embedded in DSM reconstruction and 3D models, supporting more reliable damage assessments.
Parameter optimization for creating reliable photogrammetric models in emergency scenarios
Applied Geomatics, 2018
An optimized planning and realization of the survey, coupled with well thought-out processing, allows obtaining good quality results, while guaranteeing a reasonable use of resources and time. It represents a benefit for both operators and end-users. The former can save time and acquire smaller datasets to process, while the latter can invest their resources better. These goals are even more important in case of an emergency, because the circumstances can quickly change, causing risk to both people and goods. The paper examines the possibility of using Unmanned Aerial Systems (UAS) photogrammetry for 3D modelling in such scenario, focusing on finding a compromise between the final accuracy and the requested processing time. An experimental test has been conducted over the Castle of Casalbagliano, a damaged structure located near Alessandria (Piedmont, Italy), simulating a postemergency scenario. Several processing strategies have been tested to define a workflow useful in this kind of situations. The quality of the different processing has been evaluated in terms of both residuals of the bundle block adjustment and quality of the generated dense point cloud, compared with a reference Terrestrial Laser Scanner acquisition. Finally, the possibility of publishing the obtained 3D models on the web has been exploited too.
Summer Computer Simulation Conference, 2013
Various natural and human-made events can occur in urban settings resulting in buildings collapsing and trapping victims. The task of a structural engineer is to survey the resulting rubble to assess its safety and arrange for structural stabilization, where necessary. Urban Search and Rescue (USAR) operations can then begin to locate and rescue people. Our previous work reported the use of an Unmanned Aerial Vehicle (UAV) equipped with a RGB-Depth sensor to build 3D point cloud models of disaster scenes. In this paper we extend this work by converting the point clouds into 3D models and importing them into a state-of-the-art game engine. We present a method to use these models to allow first responders to interact with the simulated rubble environment in real-time, without risk to human life. Experiments are conducted measuring traversal time both in the real world environment and using the simulation. We argue that this work will improve the safety of workers and allow a better understanding of extremely dangerous environments without unnecessary exposure during disaster response planning. * In this work we use the term "patient" in the USAR context-meaning any person trapped in rubble, in need of rescue.
Real-Time 3D Reconstruction from Images Taken from an Uav
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2015
We designed a method for creating 3D models of objects and areas from two aerial images acquired from an UAV. The models are generated automatically and in real-time, and consist in dense and true-colour reconstructions of the considered areas, which give the impression to the operator to be physically present within the scene. The proposed method only needs a cheap compact camera, mounted on a small UAV. No additional instrumentation is necessary, so that the costs are very limited. The method consists of two main parts: the design of the acquisition system and the 3D reconstruction algorithm. In the first part, the choices for the acquisition geometry and for the camera parameters are optimized, in order to yield the best performance. In the second part, a reconstruction algorithm extracts the 3D model from the two acquired images, maximizing the accuracy under the real-time constraint. A test was performed in monitoring a construction yard, obtaining very promising results. Highly realistic and easy-to-interpret 3D models of objects and areas of interest were produced in less than one second, with an accuracy of about 0.5m. For its characteristics, the designed method is suitable for video-surveillance, remote sensing and monitoring, especially in those applications that require intuitive and reliable information quickly, as disasters monitoring, search and rescue and area surveillance.
3D Applications in Disaster Mitigation and Management: Core Results of Ditac Project
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2013
According to statistical data, natural disasters as well as the number of people affected by them are occurring with increasing frequency compared to the past. This situation is also seen in Europe Union; So, Strengthening the EU capacity to respond to Disasters is very important. This paper represents the baseline results of the FP-7 founded DITAC project, which aims to develop a holistic and highly structured curriculum for responders and strategic crisis managers. Up-to-date geospatial information is required in order to create an effective disaster response plan. Common sources for geospatial information such as Google Earth, GIS databases, and aerial surveys are frequently outdated, or insufficient. This limits the effectiveness of disaster planning. Disaster Management has become an issue of growing importance. Planning for and managing large scale emergencies is complex. The number of both victims and relief workers is large and the time pressure is extreme. Emergency response and triage systems with 2D user interfaces are currently under development and evaluation. Disasters present a number of spatially related problems and an overwhelming quantity of information. 3D user interfaces are well suited for intuitively solving basic emergency response tasks. Such tasks include commanding rescue agents and prioritizing the disaster victims according to the severity of their medical condition. Further, 3D UIs hold significant potential for improving the coordination of rescuers as well as their awareness of relief workers from other organizations. This paper describes the outline of a module in a Disaster Management Course related to 3D Applications in Disaster Mitigation and Management. By doing this, the paper describes the gaps in existing systems and solutions. Satellite imageries and digital elevation data of Turkey are investigated for detecting sites prone to natural hazards. Digital image processing methods used to enhance satellite data and to produce morphometric maps in order to contribute to the detection of causal factors related to landslides, local site conditions influencing and/or experiencing earthquake damage intensity or those of tsunami and storm surge hazard sites at the coasts.
Interior models of earthquake damaged buildings for search and rescue
Advanced Engineering Informatics, 2016
Survivors trapped in void spaces formed when buildings collapse in an earthquake may be saved if search and rescue (SAR) operations are quick. A novel computational approach aims to provide building information that can guide SAR teams, thus minimizing their risk and accelerating operations. The inputs are an 'as-built' BIM model of the building before an earthquake and a partial 'as-damaged' BIM model of the exterior components after the earthquake derived from a terrestrial laser scan. A large set of possible collapse patterns is generated before the earthquake. After the event, the pattern with geometry most similar to that of the 'as-damaged' exterior BIM can be selected rapidly. This paper details the selection methods, which use least sum of point distances and Modal Assurance Criteria (MAC) algorithms, and illustrates their operation on a series of simulated computer models of collapsed structures, thus demonstrating the potential feasibility of the proposed approach.