CVIAN

该数据集是为了评估CVDisaster框架而创建的跨视角数据集，它包含了与飓风IAN相关的街景图像（SVI）和甚高分辨率（VHR）卫星图像。数据集根据视觉指标对损害感知程度进行了分类（轻微、中等、严重），这有助于灾害响应和恢复计划的制定。该数据集规模宏大，处理了957,539张街景图像，其中有1,135张图像是手动标注了损害感知程度的。其任务包括地理定位和损害感知估计。

This cross-view dataset was created to evaluate the CVDisaster framework. It contains street view images (SVI) and very high-resolution (VHR) satellite images related to Hurricane IAN. The dataset classifies the degree of perceived damage into three categories: slight, moderate, and severe, based on visual metrics, which facilitates the formulation of disaster response and recovery plans. This is a large-scale dataset, with 957,539 street view images processed, among which 1,135 images have been manually annotated with their perceived damage degrees. Its supported tasks include geolocation and perceived damage estimation.