HRSID: A High-Resolution SAR Images Dataset for Ship Detection and Instance Segmentation
HRSID: A High-Resolution SAR Images Dataset for Ship Detection and Instance Segmentation
复制标题
DOI：
10.1109/access.2020.3005861
复制
发表时间：
2020-01-01
期刊：
影响因子：
3.9
通讯作者：
Shi, Jun
中科院分区：
文献类型：
--
作者：
Wei, Shunjun;Zeng, Xiangfeng;Shi, Jun
随着卫星技术的发展，最新的合成孔径雷达（SAR）卫星成像模式可以提供更高分辨率的SAR图像，有利于舰船目标的检测和目标分割。同时，基于卷积神经网络（CNN）的目标检测器在不进行海陆分割的情况下也表现出较高的SAR舰船检测性能，但现有SAR舰船数据集存在SAR图像尺寸较小、SAR训练样本有限、标注不当等不足，阻碍了相关研究。为了促进基于CNN的舰船检测和实例分割的发展，我们构建了一个高分辨率SAR图像数据集（HRSID）。除了对象检测，实例分割也可以在HRSID上实现。在数据集构建方面，在25%的重叠率下，将136幅分辨率为1 m至5 m的全景SAR图像裁剪为800 x 800像素的SAR图像。为了减少错注和漏注，利用光学遥感影像减少港口建设的干扰。HRSID中有5604幅裁剪后的SAR图像和16951艘舰船，我们将HRSID分为训练集（65%的SAR图像）和测试集（35%的SAR图像），采用Microsoft Common Objects in Context（MS COCO）格式。8个国家的最先进的探测器在HRSID上进行实验，以建立基线; MS COCO评估指标用于综合评估。实验结果表明，HRSID可以很好地实现船舶检测和实例分割。
With the development of satellite technology, up to date imaging mode of synthetic aperture radar (SAR) satellite can provide higher resolution SAR imageries, which benefits ship detection and instance segmentation. Meanwhile, object detectors based on convolutional neural network (CNN) show high performance on SAR ship detection even without land-ocean segmentation; but with respective shortcomings, such as the relatively small size of SAR images for ship detection, limited SAR training samples, and inappropriate annotations, in existing SAR ship datasets, related research is hampered. To promote the development of CNN based ship detection and instance segmentation, we have constructed a High-Resolution SAR Images Dataset (HRSID). In addition to object detection, instance segmentation can also be implemented on HRSID. As for dataset construction, under the overlapped ratio of 25%, 136 panoramic SAR imageries with ranging resolution from 1m to 5m are cropped to 800 x 800 pixels SAR images. To reduce wrong annotation and missing annotation, optical remote sensing imageries are applied to reduce the interferes from harbor constructions. There are 5604 cropped SAR images and 16951 ships in HRSID, and we have divided HRSID into a training set (65% SAR images) and test set (35% SAR images) with the format of Microsoft Common Objects in Context (MS COCO). 8 state-of-the-art detectors are experimented on HRSID to build the baseline; MS COCO evaluation metrics are applicated for comprehensive evaluation. Experimental results reveal that ship detection and instance segmentation can be well implemented on HRSID.