A Community Data Set for Comparing Automated Coronal Hole Detection Schemes

<b>Description:</b><br>Welcome to the "Coronal Hole Detection Comparison Dataset" hosted on Figshare. The dataset comprises 29 manually selected Solar Dynamics Observatory (SDO) images, and is accompanied by the coronal hole detection results from the most widely used automated detection schemes in the solar and heliospheric science community. This community dataset was created as part of the ISWAT Coronal Hole Boundary Working Team to support the comparison and evaluation of automated coronal hole detection schemes.<b>Dataset Details:</b><b>SDO Images:</b> The dataset consists of 29 SDO images (4k x 4k) observed between the years 2014 and 2019, spanning from maximum solar activity to the subsequent minimum. We provide all seven EUV wavebands ranging from 9.4 nm to 33.5 nm captured by the AIA instrument and the line-of-sight measurements of the photospheric magnetic field from the HMI instrument for each date. These images capture a wide range of appearances of coronal holes, presenting challenging scenarios.<b>Coronal Hole Labels:</b> Each image in the dataset is accompanied by manually assigned coronal hole labels, annotated according to the criteria outlined in Reiss et al. (2023). These labels serve as the ground truth for evaluating the accuracy of automated detection schemes in terms of event-based validation, as discussed in Section 4.4 of our publication.<b>Detection Results:</b> After performing basic preprocessing steps on level 1.0 data from the SDO data platform, we provided the complete dataset to 14 participating research teams and collected the resulting coronal hole detection results, which are included here.<b>Result Collage Images:</b> A comparison of the results from all 14 different automated detection schemes is provided as a PDF file for each of the dates.<b>Intended Usage:</b><br>Researchers and data scientists who are interested in advancing the field of automated coronal hole detection are encouraged to use this dataset for benchmarking and comparing their algorithms. The dataset is intended to serve as a common ground for evaluating algorithm performance and fostering collaboration within the community.<b>Citation:</b><br>If you use this dataset in your research, please cite the following reference: Reiss et al., "<i>A Community Data Set for Comparing Automated Coronal Hole Detection Scheme</i>," ApJS, 2023.<b>Contact Information:</b>Martin A. Reiss (Community Coordinated Modeling Center, NASA Goddard, USA): [e-mail] Muglach (NASA Goddard, USA): [e-mail]<b>Disclaimer:</b><br>While the dataset has been carefully curated and annotated, we do not guarantee the accuracy or completeness of the provided labels, and any findings or conclusions drawn from the dataset should be verified.<br><br>We hope that the "Automated Coronal Hole Detection Comparison Dataset" serves as a valuable resource for advancing automated coronal hole detection schemes.<b>数据集描述：</b> 欢迎使用托管于Figshare平台的"冕洞检测对比数据集（Coronal Hole Detection Comparison Dataset）"。本数据集包含29张人工甄选的太阳动力学天文台（Solar Dynamics Observatory, SDO）图像，并附带了太阳与日球层科学领域内应用最广泛的自动化冕洞检测方案的检测结果。本社区数据集由ISWAT冕洞边界工作组打造，旨在支持自动化冕洞检测方案的对比与评估工作。 <b>数据集详情：</b> <b>SDO图像：</b> 数据集包含29张分辨率为4k×4k的SDO图像，采集时间为2014年至2019年，覆盖了太阳活动极大年至后续极小年的完整周期。针对每个观测日期，我们均提供了AIA仪器采集的9.4 nm至33.5 nm范围内的全部7个极紫外（EUV）波段数据，以及HMI仪器获取的光球磁场视向测量数据。这些图像涵盖了冕洞的多种形态特征，包含诸多具有挑战性的观测场景。 <b>冕洞标注：</b> 数据集中的每张图像均附带人工标注的冕洞标签，标注依据Reiss等人2023年提出的标准完成。正如我们发表论文的4.4章节所述，这些标签可作为基于事件验证的自动化检测方案精度评估的基准真值（ground truth）。 <b>检测结果：</b> 我们先对SDO数据平台的1.0级数据完成基础预处理，随后将完整数据集分发给14个参与的研究团队，并收集到了对应的冕洞检测结果，本数据集现已包含全部结果。 <b>结果拼接图像：</b> 针对每个观测日期，我们均提供了一份PDF文件，汇总对比了全部14种自动化检测方案的结果。 <b>适用场景：</b> 我们鼓励致力于推进自动化冕洞检测领域发展的研究人员与数据科学家，使用本数据集开展算法基准测试与对比工作。本数据集旨在作为评估算法性能、推动本社区内部协作的公共基准平台。 <b>引用方式：</b> 若您在研究中使用本数据集，请引用如下文献：Reiss等人发表于《天体物理学杂志增刊（ApJS）》2023年的《用于自动化冕洞检测方案对比的社区数据集》（*A Community Data Set for Comparing Automated Coronal Hole Detection Scheme*）。 <b>联系方式：</b> 马丁·A·赖斯（美国NASA戈达德太空飞行中心社区协调建模中心）：[e-mail] 卡琳·穆格拉赫（美国NASA戈达德太空飞行中心）：[e-mail] <b>免责声明：</b> 尽管本数据集经过精心甄选与标注，但我们不保证所提供标签的准确性与完整性，从本数据集得出的任何发现或结论均需经过验证。 我们衷心希望本冕洞检测对比数据集能够为自动化冕洞检测方案的发展提供宝贵的支持资源。