An official website of the United States government
Reconnaissance teams collect perishable data after each disaster to learn about building performance. However, often these large image sets are not adequately curated, nor do they have sufficient metadata (e.g., GPS), hindering any chance to identify images from the same building when collected by different reconnaissance teams. In this study, Siamese convolutional neural networks (S‐CNN) are implemented and repurposed to establish a building search capability suitable for post‐disaster imagery. This method can automatically rank and retrieve corresponding building images in response to a single query using an image. In the demonstration, we utilize real‐world images collected from 174 reinforced‐concrete buildings affected by the 2016 Southern Taiwan and the 2017 Pohang (South Korea) earthquake events. A quantitative performance evaluation is conducted by examining two metrics introduced for this application: Similarity Score (SS) and Similarity Rank (SR).
more »
« less
Detailed 3D point cloud data at Tumwata Village in Oregon City, OR (2021-2022):Subtitle
This Grant for Rapid Response Research (RAPID) project will collect and analyze perishable data on historical buildings. The Tumwata Village (formerly known as Blue Heron Paper Mill Site) located by the Willamette Falls in Oregon City, Oregon, has a very intriguing history and was recently purchased by the Confederated Tribes of Grand Ronde with the intent to restore the falls to their natural state and preserve some of the oldest structures. The site presents a unique opportunity to perform rapid investigations to collect and analyze perishable data on these historical buildings and develop new knowledge in the area of building assessments in corrosive environments. This industrial site contains a wide range of structure types (steel frames, concrete frames, timber frames, masonry walls and massive concrete walls) that were built over a period of 150 years and that employ many construction details that are common in older structures. The data collected and the results of the research will be applicable to many buildings in coastal communities throughout the country. Lidar data sets collected from these buildings will support the development of new methods to analyze and synthesize large data sets as well as integrate visual observations and material testing to quantify structural deterioration damages. The challenge in developing artificial intelligence (AI) technologies to find and quantify damage in structural systems using lidar data is the need to train the methods on existing data sets that show a wide range of damage states. The data to be collected from this site will provide an extensive training data set relevant to structural components common to older buildings. Development of such AI technologies for fast identification and quantification of damage would be transformative for the natural hazards research community and would expand the ability to learn from archived lidar datasets. The collected dataset will be available to researchers to serve as high quality training data in algorithm development.
more »
« less
- Award ID(s):
- 2228113
- PAR ID:
- 10544618
- Publisher / Repository:
- Designsafe-CI
- Date Published:
- Subject(s) / Keyword(s):
- Lidar Cultural Heritage 3D Modeling Seismic Analysis Finite Element Method
- Format(s):
- Medium: X
- Location:
- Corvallis, OR
- Institution:
- Oregon State University
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
At 18:08 on August 4, 2020, a large explosion occurred at Hangar 12 in the Port of Beirut. The size of the explosion was equivalent to that of an earthquake with a local magnitude (ML) of 3.3 according to the USGS. As one of the largest nonmilitary explosions to ever impact an urban region, this event provides unprecedented opportunities to document explosion impacts on urban infrastructure. To facilitate this data collection, the Geotechnical Extreme Events Reconnaissance (GEER) Association coordinated a multiagency response directed toward the collection of perishable data of engineering interest. Two main categories of infrastructure systems were impacted: the Port of Beirut and the Beirut building stock. Within the Port, the explosion triggered a quay wall failure and flow slide, and strongly impacted grain silo structures that were in close proximity to Hangar 12. Within the city, historical masonry structures, older reinforced concrete structures, and modern high-rise structures were impacted. Through a combination of in-person inspections and street-view surveys, we collected data on structural performance (including damage to load-bearing elements) and building façades. Performance levels were classified according to procedures applied following earthquakes (for structural performance) and newly proposed procedures (for façades). We describe spatial distributions of these damage types and dependencies on source distance and location-to-explosion direction. We demonstrate that physical damages correlated with damage proxy maps produced by the Jet Propulsion Laboratory and the Earth Observatory of Singapore based on Copernicus Sentinel-1 satellite synthetic aperture radar data, with a stronger correlation with structural damage than with façade damage.more » « less
-
null (Ed.)Vast amounts of damage data exist following natural disasters; the difficulty is collecting these perishable data in a timely fashion and curating them in a way that facilitates their use in follow-on research. Traditional on-the-ground reconnaissance efforts tend to focus limited human and financial resources on collecting and documenting detailed data of the most severe damage, typically in relatively small geographic areas immediately following an event. To improve post-event loss predictions, it is imperative that we collect broad and robust damage datasets, including details on good and poor performance and performance in small-to-moderate events. This presentation will introduce a framework to proactively engage local journalists and citizen scientists to collect vast amounts of real-world observational data via social media networks, such that valuable time, money, and engineering expertise can be focused on interpreting the citizen science data and on investigating the most interesting observations of damage more fully. The presentation will discuss how parts of this framework are being implemented in a National Science Foundation RAPID Response Research project to collect and curate social media data from Hurricane Florence and what future work is necessary to fully realize this citizen scientist-enabled framework for developing robust damage datasets for the natural hazards research community.more » « less
-
The 6 February 2023 Türkiye earthquakes and the accompanying aftershocks were a once-in-a-century catastrophe that has greatly impacted Türkiye and Syria. The repercussions of these events will have a lasting effect on the entire region. This article documents the geotechnical and geological observations performed by GEER (Geotechnical Extreme Events Reconnaissance) immediately following the events. Observations of ground damage, including surface fault rupture, liquefaction and lateral spreading, landslides and rock falls, and foundation failure of buildings, dams, and other civil infrastructure, are described herein. This article summarizes the key findings that were originally reported in the joint GEER-EERI (Earthquake Engineering Research Institute) reconnaissance report. The goal of these reconnaissance efforts is to document perishable data and disseminate it widely so that lessons can be learned from these events.more » « less
-
This data was gathered during the Geotechnical Extreme Events Reconnaissance (GEER) efforts following the February 6, 2023, Kahramanmaraş earthquake sequence. This dataset is comprised of terrestrial lidar scan point clouds that aim to capture liquefaction-induced building settlement, building-ground interactions, and ground deformations. The objective of the reconnaissance efforts was to capture perishable data on ground failures and liquefaction-induced infrastructure damage due to these earthquakes. Reconnaissance was performed from March 27 to April 1, 2023 in and around İskenderun, Hatay; Gölbaşı, Adıyaman; and Antakya, Hatay. Lidar scans were performed in İskenderun and Gölbaşı at selected liquefaction building sites. The reconnaissance sites were selected as those where there was evidence of liquefaction (e.g., ejecta) and liquefaction-induced building settlements, as well as building-ground interactions, and site access. The processed lidar data are included as .las point cloud files; raw data are included as .fls files. The point cloud data may be viewed and analyzed in point cloud analysis software, including the opensource software CloudCompare. Additional images of the surveyed buildings are included for reference. An explanation of the data types and structure is found in the README.pdf file. These data may be used to investigate earthquake liquefaction-induced building settlements, building-ground interactions, and liquefaction-induced ground deformations. These data will be of use and interest to engineers and researchers working in the area of liquefaction ground failures and building-ground interactions. Additional information and data from this reconnaissance are available in the GEER reports, which are referenced in the "Related Works" section.more » « less
