YESeg-OPT-SAR

Introduction

  • YESeg-OPT-SAR is a sufficiently annotated high-resolution remote sensing dataset for land use classification, which combines high-resolution RGB optical images with SAR images at a resolution of 0.5 meters.

Details

  • YESeg-OPT-SAR uses high-resolution optical and SAR images from Google Earth L18 products imaged during 2010/10/21-2016/8/25 with a spatial resolution of 0.5 m.

  • It contains 2231 pairs of images of the same area with a size of 256×256. The dataset contains two study areas, one covers an area in the longitude range of 122.836°E-122.869°E and the latitude range of 45.105°N-45.133°N, and the other with a longitude range of 110.135°E-110.160°E and latitude range of 34.593°N-34.626°N,which covering a wide range of different terrain and vegetation.The location of the study area is shown on the map below. exmple.png

  • The images in the dataset are fully labeled with pixel-level categories, including background, bare ground, vegetation, trees, houses, water, roads, and others, where the label values and category percentages are shown in the table below.

Label Name Percent(%)
0 background 2.26
1 bare ground 55.84
2 low vegetation 26.55
3 trees 5.29
4 houses 4.39
5 water 1.69
6 roads 3.88
7 other 0.11

Dataset Organization

The dataset are organized as:

├─YESeg-OPT-SAR dataset
│  ├─label
│  │  ├─1
│  │  │    1.png
│  │  │    ……
│  │  ├─10
│  │  ├─11
│  │  ├─……
│  │  ├─HY_C
│  │  └─HY_D
│  ├─optical
│  │  ├─1
│  │  ├─10
│  │  ├─11
│  │  ├─……
│  │  ├─HY_C
│  │  └─HY_D
│  └─sar
│      ├─1
│      ├─10
│      ├─11
│      ├─……
│      ├─HY_C
│      └─HY_D

Examples of Annotated Images

The optical images and SAR images of some typical features in the YESeg-OPT-SAR dataset with the corresponding annotation results are shown below. exmple.png

Contact

If you encounter any problem in using YESeg-OPT-SAR dataset or have any feedback, please contact:

Source Dataset

Dataset is avaliable:

Googledrive: https://drive.google.com/file/d/1AWgZis0ygT2UjosdIaVMqr8lNEEmhC5o/view?usp=sharing

BaiduYun: https://pan.baidu.com/s/1uBKAG4W5S-o87t4u0F3nJQ?pwd=w4x3 password:w4x3