Broad Area Damage Assessment
| Autor: | Matthew D. Reisman, Latisha Konz, Zachary J. DeSantis, Shabab E. Siddiq |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Pixel
business.industry Deep learning 02 engineering and technology Image segmentation Machine learning computer.software_genre Weighting 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Segmentation Artificial intelligence Change assessment Natural disaster business computer |
| Zdroj: | AIPR |
| DOI: | 10.1109/aipr50011.2020.9425272 |
| Popis: | Deep-learning-based broad area damage/change assessment (BADA) monitors environmental change due to disaster damage. Our approach was developed while competing in the xView2 challenge. The purpose of the xView2 challenge was to provide a capability to aid humanitarian relief efforts in response to a natural disaster. Given a pre-disaster and a post-disaster image (satellite EO), the goal of the challenge was to locate structures in the imagery and, on a per-structure basis, apply a damage level classification: no damage, minor damage, major damage, or destroyed. We extend the xBD dataset [1], [2], which is the dataset used for the xVeiw2 challenge, using a nearest neighbor label transfer and a distance-based confidence weighting, to allow our Siamese U-net to both predict damage levels at every pixel in the image and, combined with a U-net for building segmentation, provide damage assessment to individual buildings. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|