An uncertainty-aware framework for reliable disaster damage assessment via crowdsourcing

Khajwal, Asim Bashir; Noshadravan, Arash

doi:10.1016/j.ijdrr.2021.102110

Cited by 30 publications

(16 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The crowdsourced damage assessment adopted in this study is based on the methodology presented in Khajwal and Noshadravan (2021). The responses received from the participants are inferred using two approaches.…”

Section: Resultsmentioning

confidence: 99%

“…These efforts broadly fall into two categories: crowdsourcing or participatory damage assessment and automated damage assessment using artificial intelligence (AI). Many recent studies have highlighted the use of citizen science and participatory approach to enhance data collection for infrastructure monitoring (Li et al., 2021) and disaster impact assessment (Gharaibeh et al., 2021; Khajwal & Noshadravan, 2021). On the other hand, the promise of AI has not only been encouraging in solving scientific problems in other domains (Rafiei & Adeli, 2017a), it is also emerging as a promising solution for automation and enhancing the efficiency of structural health monitoring and post‐disaster damage assessment (Xu et al., 2020).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Post‐disaster damage classification based on deep multi‐view image fusion

Khajwal

Cheng

Noshadravan

2022

Computer aided Civil Eng

Self Cite

View full text Add to dashboard Cite

This study aims to facilitate a more reliable automated postdisaster assessment of damaged buildings based on the use of multiple view imagery. Toward this, a Multi‐View Convolutional Neural Network (MV‐CNN) architecture is proposed, which combines the information from different views of a damaged building, resulting in 3‐D aggregation of the 2‐D damage features from each view. This spatial 3‐D context damage information will result in more accurate and reliable damage quantification in the affected buildings. For validation, the presented model is trained and tested on a real‐world visual data set of expert‐labeled buildings following Hurricane Harvey. The developed model demonstrates an accuracy of 65% in predicting the exact damage states of buildings, and around 81% considering ±1 class deviation from ground‐truth, based on a five‐level damage scale. Value of information (VOI) analysis reveals that the hybrid models, which consider at least one aerial and ground view, perform better.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Post‐disaster damage classification based on deep multi‐view image fusion

Khajwal

Cheng

Noshadravan

2022

Computer aided Civil Eng

Self Cite

View full text Add to dashboard Cite

show abstract

“…In DoriaNET, building damage severities were merely annotated by one annotator (one of the coauthors). While the DoriaNET annotation quality was monitored and well‐controlled, 42 multiple humans may label the same building differently, particularly when the building condition does not exhibit a clear damage state, such as a common case where the damage appears to be in the range of several damage levels) 65 . In order to explore this further, we design a small‐scale experiment and conduct a comparative analysis of our uncertainty‐aware CNN predictions and collective human decisions from multiple annotators.…”

Section: Resultsmentioning

confidence: 99%

“…While the DoriaNET annotation quality was monitored and well-controlled, 42 multiple humans may label the same building differently, particularly when the building condition does not exhibit a clear damage state, such as a common case where the damage appears to be in the range of several damage levels). 65 In order to explore this further, we design a smallscale experiment and conduct a comparative analysis of our uncertainty-aware CNN predictions and collective human decisions from multiple annotators. The experiment involves 20 human annotators who assign damage labels based on the FEMA rating system to five selected building images from DoriaNET.…”

Section: Consistency Of Model Predictions and Crowdsourced Annotation...mentioning

confidence: 99%

Uncertainty‐aware convolutional neural network for explainable artificial intelligence‐assisted disaster damage assessment

Cheng

Behzadan

Noshadravan

2022

Structural Contr & Hlth

Self Cite

View full text Add to dashboard Cite

Accurate damage assessment is a critical step in post-disaster risk assessment, mitigation, and recovery. Current practices performed by experts and reconnaissance teams in the form of field evaluation require considerable time and resources. Recent advances in remote sensing imagery, artificial intelligence (AI), and computer vision have enhanced automated and rapid disaster damage assessment. Recent literature has shown promising progress in AI-assisted aerial damage assessment. However, accounting for the uncertainty in the outcome for improved quantification of confidence and enhanced model explainability for human decision-makers remains one of the key challenges.Overlooking uncertainty can lead to erroneous decisions, especially in highlyconsequential tasks such as damage assessment. The aim of this study is to develop uncertainty-aware deep learning models for the assessment of postdisaster damage using aerial imaging. Within the framework of variational Bayesian inference, Monte Carlo dropout sampling technique is used to propagate epistemic uncertainty in model predictions. With this stochastic setting, the model produces damage prediction labels with softmax as random variables, which helps quantify confidence in the model outcome using appropriate measures of uncertainty. Two networks are implemented and trained separately on two different disaster damage datasets consisting of unmanned aerial vehicle building footage as well as satellite-captured post-disaster imagery. The first network attains 59.4% accuracy in building classification, and the second network gives an accuracy of 55.1%. Results from uncertainty analysis, model confidence quantification, and analyzing model attention zone can lead to more explainable and risk-informed automated damage assessment outcomes using AI technology.

show abstract

“…Damage assessment can also be used as a consideration in making decisions. Damage Assessment contains information on the losses generated by disasters [8].…”

Section: Introductionmentioning

confidence: 99%

Flash flood (Δ) risk and damage assessment in batu, East Java

et al. 2022

View full text Add to dashboard Cite

Hydrometeorological disasters are showing an increasing trend in Indonesia. Flash floods are part of a hydrometeorological disaster that has a significant livelihood impact. Flash Flood is triggered by the intensity of extreme rain, several actions of mitigation can be taken by early warning systems, hazard and risk mapping, community preparedness, and climate change adaptation. So, how does future land use have an impact, and how much loss will result from the flash flood disaster in Batu City? The hydrometeorological disaster that occurred in Indonesia was the Batu Flash Flood. The Flash Flood occurred on November 4, 2021. The flash flood has a lot of impact on many locations, including the Bumiaji District and Junrejo District. Based on the disaster history recorded, the flash flood in Batu has only happened once, but the impact was quite big because there are many houses in the midstream of Bulukerto. Based on the losses caused, this can be the basis for efforts to control the spatial pattern of Batu City in the future. The methodology used in this study is risk assessment. While the risk study related to delta (Δ) the study of flash floods risk in spatiotemporal prevention uses the 2030 spatial plan for delta prediction (Δ), which can later predict the consequences of climate change and meteorological disasters from flash floods in Batu. The results of this study are the delta (Δ) of flash flood risk and the damage assessment of the flash flood that occurred in Batu.

show abstract

An uncertainty-aware framework for reliable disaster damage assessment via crowdsourcing

Cited by 30 publications

References 36 publications

Post‐disaster damage classification based on deep multi‐view image fusion

Post‐disaster damage classification based on deep multi‐view image fusion

Uncertainty‐aware convolutional neural network for explainable artificial intelligence‐assisted disaster damage assessment

Flash flood (Δ) risk and damage assessment in batu, East Java

Contact Info

Product

Resources

About