Post-Disaster Ruins Drone Aerial Damage Assessment Dataset

The current real estate industry faces an increasing risk of natural disasters and an urgent need for rapid post-disaster assessment and recovery. Existing assessment methods rely heavily on on-site manual inspections, which are inefficient and costly, especially posing personal safety risks in post-disaster harsh environments. The Post-Disaster Ruins Drone Aerial Damage Assessment Dataset aims to enable fast, low-cost, and safe building damage assessments through high-resolution aerial images and precise damage annotations. This dataset is collected from actual post-disaster scenarios, using advanced drone equipment under conditions without personal safety threats. The data undergo multiple rounds of annotation and consistency checks, reviewed by a professional team with architecture and engineering backgrounds to ensure annotation accuracy and consistency. Data preprocessing includes image enhancement, noise filtering, and lighting correction, and is ultimately stored in JPG format, finely organized by region and damage type. The core advantage lies in the high precision and completeness of the data quality, and the use of innovative annotation methods and data augmentation techniques significantly enhances damage recognition accuracy. In application, this dataset can significantly improve post-disaster damage assessment speed and decision-making accuracy, increasing assessment efficiency by at least 30%. Compared to other datasets, this dataset holds unique advantages in diversity and scarcity, demonstrating good scalability, and is suitable for damage assessment in different disaster scenarios.