HH-channel dataset and PolSAR dataset
收藏资源简介：
HH-channel数据集是为高分辨率SAR目标识别应用设计的，通过场景设置、飞行路径规划和雷达数据采集，收集了四个场景的高分辨率SAR目标数据。数据集涵盖了7大类21种车型，背景包括混凝土地面、草地和灌木丛。包括Ku波段和X波段数据，观测角度覆盖5个仰角和72个方位角，共形成3,308个SAR目标芯片样本。数据集提供SAR振幅图像和相应的复数值数据，可支持高分辨率SAR目标识别、目标散射特性分析及相关方法的研究。PolSAR数据集涵盖了7大类21种车型，背景包括混凝土地面、草地和灌木丛三种典型类型。包括Ku波段和X波段数据，5个仰角和最多72个方位角，提供3,308个Pauli-RGB图像和相应的极化散射矩阵数据。该数据集可为车辆目标极化散射特性分析、细粒度识别、少样本学习和域适应研究提供高质量支持，并进一步促进PolSAR目标识别技术的可靠应用。
The HH-channel dataset is designed for high-resolution SAR target recognition applications. It collects high-resolution SAR target data from four scenarios through scenario setup, flight path planning and radar data acquisition. The dataset covers 21 vehicle models under 7 major categories, with backgrounds including concrete ground, grassland and shrubs. It includes Ku-band and X-band data, with observation angles covering 5 elevation angles and 72 azimuth angles, resulting in a total of 3,308 SAR target chip samples. The dataset provides SAR amplitude images and corresponding complex-value data, which can support research on high-resolution SAR target recognition, target scattering characteristic analysis and related methods. The PolSAR dataset covers 21 vehicle models under 7 major categories, with three typical backgrounds: concrete ground, grassland and shrubs. It includes Ku-band and X-band data, with 5 elevation angles and up to 72 azimuth angles, providing 3,308 Pauli-RGB images and corresponding polarimetric scattering matrix data. This dataset can provide high-quality support for research on polarimetric scattering characteristic analysis of vehicle targets, fine-grained recognition, few-shot learning and domain adaptation, and further promote the reliable application of PolSAR target recognition technology.
数据集概述
数据集来源
本仓库为论文《Azi-Vec: High-resolution Azimuth-Sensitive PolSAR Image Generation for Vehicle Automatic Target Recognition》提供相关代码和数据集。
包含的数据集
1. HH通道数据集
- 访问地址：https://www.scidb.cn/file?fid=1d2a817c2267d47a9a5eb74ce86cc79d&mode=front
- 类型：单极化数据集
- 设计用途：面向高分辨率SAR目标识别应用。
- 数据采集：通过场景设置、飞行航迹规划与雷达数据采集，获取了四个场景的高分辨率SAR目标数据。
- 目标类别：覆盖7个大类、21种车型。
- 背景类型：包括水泥地、草地、灌木丛。
- 波段：包含Ku波段和X波段数据。
- 观测角度：覆盖5个俯仰角和72个方位角。
- 样本数量：共计3，308个SAR目标切片样本。
- 数据格式：提供SAR幅度图像及对应的复数值数据。
- 支持的研究方向：高分辨率SAR目标识别、目标散射特性分析及相关方法研究。
- 详细信息页面：https://www.sciengine.com/CSD/doi/10.11922/11-6035.csd.2025.0094.zh
2. 极化合成孔径雷达数据集
- 访问地址：https://www.scidb.cn/detail?dataSetId=0378a4e05fdf43c4b81187e4448e369c&version=V1
- 类型：PolSAR数据集
- 目标类别：覆盖7个大类、21种车型。
- 背景类型：三种典型背景，包括水泥地、草地、灌木丛。
- 波段：包含Ku波段和X波段数据。
- 观测角度：5个俯仰角，最多72个方位角。
- 数据内容：提供共计3，308幅Pauli-RGB图像及对应的极化散射矩阵数据。
- 支持的研究方向：车辆目标极化散射特性分析、细粒度识别、少样本学习、域自适应等研究，可进一步促进PolSAR目标识别技术的可靠应用。
引用信息
若使用上述两个数据集，请引用以下文献：
bibtex @article{Wang2025HRSARVehicleDataset, author = {Wei Wang and Huiqiang Zhang and Hongqi Fan and Zhongjie Wu and Bingwei Hui and Tianpeng Liu and Shujie Wu and Deliang Xiang and Chaofu Wei and Yingbing Ma}, title = {A dataset of high-resolution SAR vehicle target recognition}, journal = {China Scientific Data}, year = {2025}, doi = {10.11922/11-6035.csd.2025.0094.zh}, url = {https://www.sciengine.com/CSD/doi/10.11922/11-6035.csd.2025.0094.zh} }
@misc{0378a4e05fdf43c4b81187e4448e369c, author = {Wang Wei and Deng Jie and Fan Hongqi and Wu Zhongjie and Quan Sinong and Hui Bingwei and Liu Tianpen and Wu Shujie and Xiang Deliang and Wei Chaofu}, title = {{High-resolution Airborne PolSAR vehicle target recognition dataset}}, year = 2025, month = nov, publisher = {Science Data Bank}, version = {V1}, doi = {10.57760/sciencedb.j00001.01627}, url = https://doi.org/10.57760/sciencedb.j00001.01627 }
