flood-binary-hnm-benchmark
收藏资源简介:
该数据集是一个用于洪水检测的图像分类数据集,包含4,154个训练样本,总大小约7.64GB。每个样本包含以下特征:1) 图像数据(image字段);2) 二进制分类标签(label字段),0表示'flood'(洪水),1表示'non_flood'(非洪水);3) 类别信息(category字符串字段);4) 数据来源(source字符串字段)。数据集仅包含训练集划分,数据文件存储于'train-*'路径下。该数据集适用于计算机视觉领域的二元分类任务,特别是洪水检测相关的应用场景。
This is an image classification dataset for flood detection, consisting of 4,154 training samples with a total size of approximately 7.64 GB. Each sample contains the following features: 1) Image data (stored in the `image` field); 2) Binary classification label (stored in the `label` field), where 0 represents 'flood' and 1 represents 'non_flood'; 3) Category information (a string field named `category`); 4) Data source (a string field named `source`). The dataset only contains a single training split, with all data files stored under paths matching the pattern 'train-*'. This dataset is suitable for binary classification tasks in the field of computer vision, especially for flood detection-related application scenarios.
数据集概述
基本信息
- 数据集名称: flood-binary-hnm-benchmark
- 发布者: zinnia82
- 存储库地址: https://huggingface.co/datasets/zinnia82/flood-binary-hnm-benchmark
数据集结构与内容
- 数据格式: 图像分类数据集
- 特征字段:
image: 图像数据,格式为imagelabel: 分类标签,为类别标签,包含两个类别:0代表flood(洪水),1代表non_flood(非洪水)category: 类别信息,数据类型为字符串source: 数据来源,数据类型为字符串
数据规模与划分
- 唯一数据划分: train(训练集)
- 训练集样本数量: 4154
- 训练集大小: 7635455737字节
- 下载大小: 7387727042字节
- 数据集总大小: 7635455737字节
配置与访问
- 默认配置名称: default
- 数据文件路径: data/train-*




