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

Abstract

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.

Disaster Image Classification by Fusing Multimodal Social Media Data

“…However, our target was to explore a method that is able to complete damage assessment tasks with high accuracy with minimal data. For example, generative adversarial networks (GAN) can be explored to assess the severity of disaster damage [40]. In future work, we will focus on addressing this open problem.…”

Generative Adversarial Networks for single image with high quality image

“…This enables the network to generate high quality and realistic images. Tilon et al used the generative adversarial network to observe the image and detect the damage of buildings after the disaster [16]. Because this method can be transferred to different types of damage or geographical locations, it has stronger applicability.…”

A Framework to Model the Wind-Induced Losses in Buildings during Hurricanes

“…Dollar losses in the year 2020 related to hurricanes were close to $20 billion, and totaled $92 billion in 2017 [4]. Recently, severe storms such as Hurricanes Ida, Irma, Maria, Michael and Matthew have caused extensive damage to insured properties [5][6][7]. Post-damage surveys [5][6][7][8][9][10] consistently report roof damage [7][8][9][10] to buildings.…”

“…Recently, severe storms such as Hurricanes Ida, Irma, Maria, Michael and Matthew have caused extensive damage to insured properties [5][6][7]. Post-damage surveys [5][6][7][8][9][10] consistently report roof damage [7][8][9][10] to buildings. Hence, it is important to understand the various factors that influence losses during hurricane events.…”