遥感卫星图像船舶检测样本数据集研究综述

doi:10.11871/jfdc.issn.2096-742X.2026.02.016

摘要/Abstract

摘要：

【目的】 系统梳理卫星遥感图像船舶检测样本数据集的研究进展和应用现状。【文献范围】本文分析了2017—2024年文献发布的29个遥感图像船舶数据集。【方法】 从技术角度，按照遥感卫星成像载荷所采用不同波段，对可见光图像船舶检测数据集、SAR图像船舶检测数据集、红外图像船舶检测数据集，从船舶类别、船舶数量、标注方式、分辨率、数据来源、获取方式等方面进行了详细论述。在应用角度下，针对数据集特点、元数据、检测效果等三个方面进行了归纳分析。【结果】 现有数据集在船舶检测应用中发挥了重要作用，但存在着样本类别不均衡、场景不够丰富、分辨率不够高和数据规模不够大等问题，后续数据集朝着高质量、大规模、多模态融合等方向发展。本文能够为后续开展基于深度学习方法的遥感卫星图像船舶样本数据集检测与识别研究提供有价值的数据参考和指导。

关键词: 深度学习, 数据集, 船舶检测, 精确识别, 复杂场景

Abstract:

[Objective] This paper aims to systematically review the research progress and application status of ship detection sample datasets in satellite remote sensing images. [Literature Scope] A total of 29 remote sensing image ship datasets published in the literature from 2017 to 2024 are analyzed. [Methods] From a technical perspective, according to the different bands adopted by remote sensing satellite imaging payloads, the ship detection datasets of visible light images, SAR images, and infrared images are discussed in detail in terms of ship categories, number of ships, annotation methods, resolution, data sources, and acquisition methods. From an application perspective, a summary analysis is conducted focusing on three aspects: dataset characteristics, metadata, and detection effects. [Results] Existing datasets have played an important role in ship detection applications, but there are problems such as unbalanced sample categories, insufficiently rich scenes, low resolution, and small data scale. Future datasets will develop towards high quality, large scale, and multi-modal fusion. This paper can provide valuable data references and guidance for subsequent research on the detection and recognition of ship sample datasets in remote sensing satellite images based on deep learning methods.

Key words: deep learning, datasets, ship detection, accurate identification, complex backgrounds

周玉明, 张一鸣, 刘圆圆, 黄山. 遥感卫星图像船舶检测样本数据集研究综述[J]. 数据与计算发展前沿, 2026, 8(2): 215-226.

ZHOU Yuming, ZHANG Yiming, LIU Yuanyuan, HUANG Shan. A Review of the Research on Remote Sensing Satellite Image Ship Detection Sample Dataset[J]. Frontiers of Data and Computing, 2026, 8(2): 215-226, https://cstr.cn/32002.14.jfdc.CN10-1649/TP.2026.02.016.

图/表 7

表1

遥感图像船舶数据集相关信息"

序号	数据集	公开时间	图像来源及特点	分辨率	样本数量	用途	分类数量
1	ShipRSImageNet	2021	xView、HRSC2016、FGSD、空客船舶检测比挑战赛数据集；规模大、分类细、场景多；船舶个体识别支撑还不够	0.2~6	17,573	斜框检测识别	50
2	FGSC-23	2021	GF-2、HRSC2016；类别、图像和标签具有多样性；样本不均衡	0.4~2	4,052	切片识别	23
3	FGSCR-42	2021	NWPU VHR-10、RSOD、MSTAR、RSC11等数据集；更适合细粒度船舶分类；规模还需扩充	0.4~2	9,320	水平框检测识别	42
4	IceRegionShip	2024	Gaofen-1、Landsat-8和Sentinel-2；首个针对海冰区域船舶检测数据集；实例数量较小	2~16	11,436	水平框检测	1
5	DMFGRS	2024	商业卫星图像；为多模态图像数据集；场景不够丰富	0.5~3.2	3,689	斜框检测识别	20
6	LS-SSDD-v1.0	2021	Sentinel-1；图像尺寸大、小型船舶丰富；场景不够丰富	5×20	9,000	水平框检测	1
7	SRSDD-v1.0	2021	Gaofen-3，分辨率高、场景复杂；规模不大，样本类别不平衡	1	666	斜框检测识别	6
8	RSDD-SAR	2022	Gaofen-3、TerraSAR-X；多种成像模式、多种极化方式、多种分辨率；泛化能力距实际应用有差距	2~20	7,000	斜框检测	1
9	FPSD	2023	AIRSAR、UAVSAR和RS-2；场景丰富，船舶对象较小；泛化能力有限	4.6~9.2×3.3~6.6	853	水平框检测	1
10	SDFSD-v1.0	2024	GF-3、ICEYE、Capella和Umbra；涵盖主要港口和航线，背景多复杂，注释多样；存在未知类别，样本类别不平衡	0.3~1	35,787	斜框检测识别	15
11	TISD	2022	SDGSAT-1 TIS；卫星红外船数据集有效补充；分辨率、数据规模有待提升	10	12,774	水平框检测	1
12	ISDD	2023	Landsat8；场景多样、天气条件多样；船舶规模小	10~20	3,061	水平框检测	1
13	NUDT SIRST Sea	2023	规模大、背景复杂、对象类型多、对象小、多尺度；样本类别不平衡	/	17,598	水平框检测	1

表1

图1

图2

图3

图4

图5

表2

参考文献 24

[1]	王文秀. 红外遥感图像舰船目标在线检测关键技术研究[D]. 上海: 中国科学院大学(中国科学院上海技术物理研究所), 2019: 1-3.
[2]	张兰丹. 基于张量恢复的红外弱小目标检测方法研究[D]. 成都: 电子科技大学, 2020: 1-5.
[3]	XIA, G S, BAI X, DING J, et al. DOTA: A large-scale dataset for object detection in aerial images[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, San Francisco, CA, USA: IEEE, 2018: 3974-3983.
[4]	LIU Z, WANG H, WENG L, et al. Ship rotated bounding box space for ship extraction from high-resolution optical satellite images with complex backgrounds[J]. IEEE Geoscience and Remote Sensing Letters. 2016, 13(8), 1074-1078. doi: 10.1109/LGRS.2016.2565705
[5]	CHENG G, ZHOU P, HAN J. Learning rotation-invariant convolutional neural networks for object detection in VHR optical remote sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing. 2016, 54(12): 7405-7415. doi: 10.1109/TGRS.2016.2601622
[6]	ZHNAG Z, ZHANG L, WANG Y, et al. ShipRSImageNet: A Large-Scale Fine-Grained Dataset for Ship Detection in High-Resolution Optical Remote Sensing Images[J]. Applied Earth Observations and Remote Sensing. 2021, 14: 8458-8471.
[7]	LAM D, KUZMA R, MCGEE K, et al. xView: Objects in Context in Overhead Imagery[J]. 2018.DOI:10.48550/arXiv.1802.07856.
[8]	CHEN K, WU M, LIU J, et al. FGSD: A Dataset for Fine-Grained Ship Detection in High Resolution Satellite Images[J]. arXiv preprint arXiv:2003.06887. 2020: 1-8.
[9]	姚力波, 张筱晗, 吕亚飞, 等. FGSC-23: 面向深度学习精细识别的高分辨率光学遥感图像舰船目标数据集[J]. 中国图象图形学报, 2021, 26(10): 2337-2345.
[10]	DI Y H, JIANG Z G, ZHANG H P. A public dataset for fine-grained ship classification in optical remote sensing images[J]. Remote Sensing, 2021, 13(4): 747. doi: 10.3390/rs13040747
[11]	WANG P, LIU B, LI Y, et al. IceRegionShip: Optical Remote Sensing Dataset for Ship Detection in Ice-Infested Waters[J]. Applied Earth Observations and Remote Sensing, 2024, 17: 1007-1020.
[12]	SONG S, ZHANG R, HU M, et al. Fine-Grained Ship Recognition Based on Visible and Near-Infrared Multimodal Remote Sensing Images: Dataset, Methodology and Evaluation[J]. Computers, Materials & Continua. 2024, 79(3): 5244-5271.
[13]	WANG W, DENG Y, LIU D, et al. Forthcoming Spaceborne SAR Development[J]. Journal of Radars, 2020, 9(1): 1-33.
[14]	LI J, QU C, SHAO J. Ship detection in SAR images based on an improved Faster R-CNN[C]// Proceedings of the 2017 SAR in Big Data Era:Models, Methods and Applications (BIGSARDATA). Beijing, China: IEEE, 2017: 1-6.
[15]	WANG Y, WANG C, ZHANG H, et al. A SAR dataset of ship detection for deep learning under complex backgrounds[J]. Remote Sensing, 2019, 11(7): 765. doi: 10.3390/rs11070765
[16]	ZHANG T, ZHANG X, KE X, et al. LS-SSDD-v1.0: A deep learning dataset dedicated to small ship detection from large-scale Sentinel-1 SAR images[J]. Remote Sensing, 2020, 12(18): 2997. doi: 10.3390/rs12182997
[17]	LEI S, LU D, QIU X, et al. SRSDDv1.0: A high-resolution SAR rotation ship detection dataset[J]. Remote Sensing, 2021, 13(24): 5104. doi: 10.3390/rs13245104
[18]	徐从安, 苏航, 李健伟, 等. RSDD-SAR:SAR舰船斜框检测数据集[J]. 雷达学报, 2022, 11(4): 581-599.
[19]	YANG Z, ZHANG P, WANG N, et al. A Lightweight Theory-Driven Network and Its Validation on Public Fully Polarized Ship Detection Dataset[J]. Applied Earth Observations and Remote Sensin, 2024, 17:3755
[20]	CAI P, LIU B, WANG P, et al. SDFSD-v1.0: A Sub-Meter SAR Dataset for Fine-Grained Ship Detection[J]. Remote Sensing, 2024, 16(21): 3952. doi: 10.3390/rs16213952
[21]	WANG N, LI B, WEI X, et al. Ship detection in spaceborne infrared image based on lightweight CNN and multisource feature cascade decision[J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 59(5): 4324-4339. doi: 10.1109/TGRS.2020.3008993
[22]	LI L, YU J, CHEN F. TISD: A Three Bands Thermal Infrared Dataset for All Day Ship Detection in Spaceborne Imagery[J]. Remote Sens, 2022, 14(21): 5297. doi: 10.3390/rs14215297
[23]	HAN Y, LIAO J, LU T, et al. KCPNet: Knowledge-driven context perception networks for ship detection in infrared imagery[J], IEEE Transactions on Geoscience and Remote Sensing, 2023, 61.
[24]	WU T H, LI B Y, LUO Y H, et al, MTU-Net: Multilevel TransUnet for space-based infrared tiny ship detection[J], IEEE Transactions on Geoscience and Remote Sensing, 2023, 61.

序号	数据集		均值（不同模型）			方差（不同模型）			指标
序号	数据集		近岸	离岸	混合场景	近岸	离岸	混合场景	指标
1	可见光	FGSC-23	/	/	83.52%	/	/	0.06%	准确率
2		FGSCR-42	/	/	88.77%	/	/	0.23%	准确率
3		IceRegionShip	/	/	79.90%	/	/	0.15%	AP50
4	SAR	LS-SSDD-v1.0	45.17%	83.40%	70.83%	2.03%	1.02%	1.14%	F1
5		SRSDD-v1.0	39.88%	53.51%	42.30%	2.06%	3.41%	2.25%	F1
6		RSDD-SAR	61.27%	92.09%	87.80%	0.25%	0.03%	0.03%	AP50
7		FPSD	/	/	86.38%	/	/	0.22%	AP50
8		SDFSD-v1.0	/	/	56.33%	/	/	0.72%	AP50
9	红外	TISD	/	/	65.88%	/	/	0.17%	AP50
10	红外	ISDD	/	/	91.09%	/	/	0.01%	AP50