混合极化实验与分析数据集

数据主编：王宇; 陆萍萍 数据集简介：HEAD-1.0是面向混合极化（HP）微波成像雷达（也称 Synthetic Aperture Radar, SAR）研究与推广的数据集，旨在推动该模式的极化定标、散射特性分析和地物分类等核心技术的发展与突破。相比全极化（QP）SAR，HP-SAR具有更大的观测分辨率、幅宽、视场范围和更优的模糊性能同时兼顾了与之相当的地物分类能力。因此，提供可用于QP模式对比分析的数据条件，是本数据集的重要目标。数据集主要由陆地探测一号条带模式的L1级单视复数影像（SLC）构成，其中HP模式标称分辨率为3 m、成像幅宽为50 km，包括LH和LV（或RH和RV）两种极化方式；QP模式标称分辨率为6 m、成像幅宽为30 km，包括HH、HV、VH和VV四种极化方式。为满足多样化极化应用需求，HEAD-1.0还提供了地理编码的单视复数影像（GSLC）。本数据集覆盖三个具有显著地理与地貌差异的典型区域：新疆天山山脉东部的山地—戈壁过渡带、内蒙古苏尼特地区的草原—荒漠草原交错带，以及太行山及其南麓的丘陵—平原农业区。数据集面向多元极化研究需求，主要支持以下研究方向： 混合极化定标：提供定标器布设及散射矩阵等信息。利用新疆哈密与内蒙古苏尼特定标场的HP与QP影像，开发并验证HP定标方法； 地物分类：提供局部地面真值图及辅助光学影像。发展适用于HP数据的地物分类算法，并系统比较QP与HP模式在不同地物分类任务中的性能差异，明确HP模式的分类能力边界，为后续SAR任务极化模式的选择提供科学依据； 类行星模拟：提供类行星地貌标注。对比分析不同类行星地貌的HP特征，发展自动识别与提取算法，服务于行星地质学类比研究。 详细使用说明请参考：星载混合极化成像雷达评估与分析数据集（HEAD-1.0）下载使用说明.pdf

Data Editors: Wang Yu, Lu Pingping
Dataset Introduction: HEAD-1.0 is a dataset dedicated to the research and popularization of hybrid polarization (HP) microwave imaging radar, also known as Synthetic Aperture Radar (SAR). It aims to promote the development and breakthroughs of core technologies including polarization calibration, scattering characteristic analysis, and land cover classification for this SAR mode.
Compared with fully polarimetric (QP) SAR, HP-SAR offers larger observation resolution, swath width, and field of view, as well as superior ambiguity performance, while maintaining comparable land cover classification capabilities. Therefore, providing data conditions for comparative analysis with the QP mode is a key objective of this dataset.
The dataset is primarily composed of L1-level single-look complex (SLC) images acquired via the strip mode of China's Land Exploration Satellite-1. For the HP mode, the nominal resolution is 3 m, the imaging swath width is 50 km, and two polarization configurations are included: LH and LV (or RH and RV). For the QP mode, the nominal resolution is 6 m, the imaging swath width is 30 km, and four polarization configurations are supported: HH, HV, VH, and VV.
To meet diverse polarization application requirements, HEAD-1.0 also provides geocoded single-look complex (GSLC) images.
This dataset covers three typical regions with distinct geographical and geomorphic differences: the mountain-gobi transition zone in the eastern Tianshan Mountains of Xinjiang, the grassland-desert steppe ecotone in the Sonid area of Inner Mongolia, and the hilly-plain agricultural region in the Taihang Mountains and their southern foothills.
Targeted at multi-polarization research needs, this dataset mainly supports the following research directions:
1. Hybrid Polarization Calibration: Information such as calibrator layout and scattering matrices will be provided. HP and QP images from the Hami calibration site in Xinjiang and the Sonid calibration site in Inner Mongolia can be used to develop and validate HP calibration methods;
2. Land Cover Classification: Local ground truth maps and auxiliary optical images are provided. Researchers can develop land cover classification algorithms applicable to HP data, systematically compare the performance differences between QP and HP modes across various land cover classification tasks, clarify the classification capability boundaries of the HP mode, and provide scientific evidence for the selection of polarization modes for future SAR tasks;
3. Exoplanetary Analogue Simulation: Exoplanetary landscape annotations are provided. Comparative analysis of HP characteristics of different exoplanetary analog landscapes can be conducted, and automatic recognition and extraction algorithms can be developed to support planetary geology analog research.
For detailed usage instructions, please refer to the document "Download and Usage Instructions for Spaceborne Hybrid Polarization Imaging Radar Evaluation and Analysis Dataset (HEAD-1.0).pdf"