satellite-image-deep-learning/VHR-10

---
license: mit
task_categories:
- image-segmentation
- object-detection
language:
- en
tags:
- satellite
size_categories:
- 1K<n<10K
---
The VHR-10 dataset mirrored from https://github.com/chaozhong2010/VHR-10_dataset_coco

NWPU VHR-10 data set is a challenging ten-class geospatial object detection data set. This dataset contains a total of 800 VHR optical remote sensing images, where 715 color images were acquired from Google Earth with the spatial resolution ranging from 0.5 to 2 m, and 85 pansharpened color infrared images were acquired from Vaihingen data with a spatial resolution of 0.08 m. The data set is divided into two sets: a) Positive image set which contains at least one target in an image contains 650 images. b) Negative image set contains 150 images and it does not contain any targets. From this the positive image set, 757 airplanes, 302 ships, 655 storage tanks, 390 baseball diamonds, 524 tennis courts, 159 basketball courts, 163 ground track fields, 224 harbors, 124 bridges, and 477 vehicles were manually annotated with bounding boxes and instance masks used for ground truth.


原始信息汇总

NWPU VHR-10 数据集概述

基本信息

• 许可证：MIT
• 任务类别：图像分割、目标检测
• 语言：英语
• 标签：卫星
• 数据量：1K<n<10K

数据集描述

NWPU VHR-10 数据集是一个具有挑战性的十类别地理空间目标检测数据集。该数据集包含总共 800 张高分辨率光学遥感图像，其中 715 张彩色图像来自 Google Earth，空间分辨率范围从 0.5 米到 2 米，85 张融合彩色红外图像来自 Vaihingen 数据，空间分辨率为 0.08 米。

数据集分为两部分：

• 正图像集：包含至少一个目标的图像，共 650 张。
• 负图像集：不包含任何目标的图像，共 150 张。

从正图像集中，手动标注了以下目标的边界框和实例掩码作为地面真实数据：

• 飞机：757 个
• 船只：302 个
• 储罐：655 个
• 棒球场：390 个
• 网球场：524 个
• 篮球场：159 个
• 地面轨道场：163 个
• 港口：224 个
• 桥梁：124 个
• 车辆：477 个