Post-Disaster Electrical Infrastructure Damage Assessment: Satellite Imagery and AI Model Made Available

Within the scope of the project supported by the Energy Market Regulatory Authority (EPDK) of the Republic of Turkey and carried out by Istanbul Technical University in cooperation with BEDAŞ, AYEDAŞ, SEDAŞ, TREDAŞ and UEDAŞ, new generation artificial intelligence-based models were developed for damage detection of electrical infrastructure after disasters. The study aims to quickly and accurately detect damages in the electrical distribution infrastructure after natural disasters by using optical satellite images and Sentinel-1 SAR data at different resolutions.

The artificial intelligence models developed within the scope of the project and published as open source are as follows:

🔹 KATE-CD (Change Detection): Detects damage and changes by comparing pre- and post-disaster satellite images.

🔹 KATE-PD (Pre-Disaster): Performs vulnerability analysis of the electrical infrastructure before the disaster and reveals potential risks.

📦 Open Access Datasets and Models

The developed models and associated data sets have been made available to researchers, public institutions and the private sector. Through these resources, users can develop AI-supported solutions in disaster management, electrical infrastructure resilience and damage assessment studies.

🔗 Models and Codes on GitHub:

🔗 Datasets from Hugging Face:

🔗 Executable Environments via Code Ocean:

İlginizi Çekebilir

Call For Papers – IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing – Special Issue on “Damage/Risk Assessment for 2023 Turkey Earthquakes using Remote Sensing”

On February 6, two major earthquakes of 7.4 and 7.8 in magnitude hit city of …

CSCRS