星载SAR机场检测数据集

数据主编：张帆 马飞 周勇胜 星载SAR机场检测数据集，主要采用了欧空局Sentinel-1B卫星数据，构建一套包含多种极化方式、多种尺寸大小、涵盖多个国家多个城市的SAR机场目标数据集，推动SAR机场目标检测等先进技术的深入研究。数据集由北京化工大学遥感技术研究所团队张帆、马飞、周勇胜等构建。 图1 数据集样例 星载SAR机场检测数据集主要由images和labels两部分组成，images文件夹总计624张切片，包含训练集和测试集；labels文件夹包含了所有图片对应的标注文件。两个文件夹中的文件命名方式均为“国家_地区_序列号”。标注示例为图1所示，图1(a)为切片示例，图1(b)为对应的.txt标签文件，一行标注一个目标，分别记录检测类型(0:机场)、归一化的机场中心位置（列、行标号）、归一化的机场目标框宽度和归一化的机场目标框长度，符合Yolo系列、PolarMask、SSD和Faster-RCNN等主流检测网络的格式要求。 星载SAR机场检测数据集所有权归北京化工大学遥感技术研究所所有，《雷达学报》编辑部具有编辑出版权等。 读者可免费使用该数据进行教学、科研等，但需在论文、报告等成果中引用或致谢。该数据禁止私自用于商业目的，如有商业需求，请与《雷达学报》编辑部联系。 数据集使用说明见“星载SAR机场检测数据集使用说明.pdf” 致谢: 北京化工大学信息科学与技术学院遥感技术研究所王道昌、陈龙、刘颖冰等对本数据集制作作出了贡献。

Dataset Curators: Zhang Fan, Ma Fei, Zhou Yongsheng. The spaceborne SAR airport detection dataset is primarily constructed using Sentinel-1B satellite data from the European Space Agency (ESA). It is a SAR airport target dataset covering multiple polarization modes, various target sizes, multiple countries and cities, aiming to promote in-depth research on advanced technologies such as SAR airport target detection. The dataset was developed by the team of the Remote Sensing Technology Institute of Beijing University of Chemical Technology, including Zhang Fan, Ma Fei, Zhou Yongsheng and other members.

Figure 1: Dataset Samples. The spaceborne SAR airport detection dataset mainly consists of two folders: `images` and `labels`. The `images` folder contains a total of 624 slices, including training and test sets; the `labels` folder contains annotation files corresponding to all images. The file naming convention for both folders follows the format "Country_Region_SerialNumber".

The annotation examples are shown in Figure 1: Figure 1(a) is a slice sample, and Figure 1(b) is the corresponding .txt label file. Each line annotates one target, recording the detection type (0: airport), normalized central coordinates of the airport (column index, row index), normalized width and normalized length of the airport bounding box, which conforms to the format requirements of mainstream detection networks such as YOLO series, PolarMask, SSD, and Faster-RCNN.

The ownership of the spaceborne SAR airport detection dataset belongs to the Remote Sensing Technology Institute of Beijing University of Chemical Technology, and the Editorial Department of *Journal of Radars* holds the editing and publishing rights, among others.

Readers may use this dataset for teaching, research and other non-commercial purposes free of charge, but must cite or acknowledge it in their papers, reports and other academic achievements. Unauthorized commercial use of this dataset is prohibited. For commercial usage requests, please contact the Editorial Department of *Journal of Radars*.

Please refer to "Spaceborne SAR Airport Detection Dataset Usage Instructions.pdf" for detailed dataset usage guidelines.

Acknowledgements: Wang Daochang, Chen Long, Liu Yingbing and other members from the Remote Sensing Technology Institute, School of Information Science and Technology, Beijing University of Chemical Technology contributed to the development of this dataset.