WSDataset
收藏资源简介:
该仓库包含用于鲸鱼检测的卫星图像样本数据,采用YOLO格式。数据来源于Cubaynes和Fretwell的论文,覆盖了座头鲸、长须鲸、灰鲸和南露脊鲸等多种鲸鱼物种,卫星图像来源包括WorldView-3、WorldView-2、GeoEye-1和Quickbird-2。数据集结构包括训练、验证和测试集的图像和标注文件,图像格式为PNG,标注格式为TXT(YOLO格式)。数据集遵循CC BY 4.0许可。
This repository contains sample satellite image datasets for whale detection, formatted in accordance with the YOLO annotation standard. The dataset is sourced from the peer-reviewed papers authored by Cubaynes and Fretwell, and covers multiple whale species including humpback whales, fin whales, gray whales, and southern right whales. The satellite imagery used originates from platforms including WorldView-3, WorldView-2, GeoEye-1, and Quickbird-2. The dataset is organized into training, validation, and test splits, with corresponding image and annotation files for each split: all images are stored in PNG format, while annotations are saved in TXT files complying with the YOLO specification. This dataset is licensed under CC BY 4.0.
WSDataset (Sample) 数据集概述
数据集基本信息
- 数据集名称:WSDataset (Sample)
- 核心内容:用于从卫星图像中检测鲸鱼的YOLO格式样本数据。
- 重要说明:此仓库仅包含用于演示和学习目的的样本数据。
数据来源与背景
- 来源论文:Cubaynes, H.C., Fretwell, P.T. "Whales from space dataset, an annotated satellite image dataset of whales for training machine learning models." Scientific Data 9, 245 (2022).
- 原始数据集物种覆盖:座头鲸、长须鲸、灰鲸、南露脊鲸。
- 原始卫星图像来源:WorldView-3, WorldView-2, GeoEye-1, Quickbird-2。
- 许可证:遵循原始数据集的 CC BY 4.0 许可。
数据集结构
数据集目录结构如下:
WS/ ├── images/ │ ├── train/ # 训练样本 (14张图像) │ ├── val/ # 验证样本 (2张图像) │ └── test/ # 测试样本 (4张图像) └── labels/ ├── train/ # 训练标注 (14个文件) ├── val/ # 验证标注 (2个文件) └── test/ # 测试标注 (4个文件)
文件格式与内容
- 图像格式:PNG。
- 标注格式:TXT (YOLO格式)。
- 坐标格式:归一化边界框 (class x_center y_center width height)。所有坐标值归一化到 [0, 1] 范围。
- 样本文件位置:
images/train/- 训练样本图像。images/val/- 验证样本图像。images/test/- 测试样本图像。labels/- 对应的YOLO格式标注文件。
使用方式
YOLO格式训练
可用于熟悉YOLO模型训练流程,示例data.yaml配置如下:
yaml
train: ./images/train
val: ./images/val
test: ./images/test
nc: 1 # 类别数量
names: [whale] # 类别名称
标注文件格式
每个标注文件(.txt)包含图像中所有目标的边界框信息,每行格式为:
class_id x_center y_center width height
使用要求与许可
- 署名:使用时需引用原始论文。
- 学术用途:推荐仅用于学术研究和学习。
- 版权:原始卫星图像的版权归数据提供商所有。
- 许可证:CC BY 4.0。
引用
若使用此数据集,请同时引用原始论文和此仓库: bibtex @article{cubaynes2022whales, title={Whales from space dataset, an annotated satellite image dataset of whales for training machine learning models}, author={Cubaynes, Hannah C and Fretwell, Peter T}, journal={Scientific Data}, volume={9}, number={1}, pages={245}, year={2022}, publisher={Nature Publishing Group UK London}, doi={10.1038/s41597-022-01377-4} }
bibtex @article{WANG2025130778, title = {Whale Identification and Size Estimation in Satellite Imagery via Intelligent Subtle Perception}, journal = {Expert Systems with Applications}, pages = {130778}, year = {2025}, issn = {0957-4174}, doi = {https://doi.org/10.1016/j.eswa.2025.130778}, url = {https://www.sciencedirect.com/science/article/pii/S0957417425043933}, author = {Siqi Wang and Baoxiang Huang and Milena Radenkovic and Ge Chen}, }
免责声明
此仓库包含个人整理的样本数据,仅用于学习和演示。如需使用完整数据集,请参考原始论文中的数据可用性声明。




