MultiFloodSynth
收藏资源简介:
MultiFloodSynth是由韩国中央大学提出的一个合成数据集,旨在为洪水灾害检测提供高质量的训练数据。该数据集通过模拟真实世界的洪水环境,生成具有不同洪水级别的图像,并提供了包括法线图、分割图、3D边界框等在内的多种注释类型。数据集总共包含70,117张图像,分为洪水和非洪水两大类,为洪水级别检测、物体定位等计算机视觉任务提供了丰富的训练样本。
MultiFloodSynth is a synthetic dataset proposed by Chung-Ang University in South Korea, developed to provide high-quality training data for flood disaster detection. This dataset generates images with varying flood severity levels by simulating real-world flood environments, and offers multiple annotation types including normal maps, segmentation masks, 3D bounding boxes and more. In total, the dataset contains 70,117 images categorized into two classes: flood and non-flood, providing rich training samples for computer vision tasks such as flood severity detection and object localization.

- 1MultiFloodSynth: Multi-Annotated Flood Synthetic Dataset Generation韩国中央大学 · 2025年



