D1D-UNet：侧扫声呐实时海底检测的双通道一维U-Net模型

doi:10.13474/j.cnki.11-2246.2026.1017

测绘通报 ›› 2026, Vol. 0 ›› Issue (10): 112-118,186.doi: 10.13474/j.cnki.11-2246.2026.1017

• 学术研究 • 上一篇

D1D-UNet：侧扫声呐实时海底检测的双通道一维U-Net模型

严俊¹, 李洪¹, 孟俊霞², 陈书祥¹

1. 安徽大学资源与环境工程学院, 安徽合肥 230601;
2. 安徽建筑大学土木工程学院, 安徽合肥 230601

收稿日期:2026-01-16 发布日期:2026-11-06
作者简介:严俊(1989—),男,博士,副教授,主要研究方向为海洋测绘、声呐数据处理、深度学习。E-mail:jun.yan@ahu.edu.cn
基金资助:
国家自然科学基金(42104036；41906168)；安徽省自然科学基金(2308085MD124；1908085QD161)

D1D-UNet: a dual-channel one-dimensional U-Net model for real-time bottom detection with side scan sonar

Yan Jun¹, Li Hong¹, Meng Junxia², Chen Shuxiang¹

1. School of Resources and Environmental Engineering, Anhui University, Hefei 230601, China;
2. School of Civil Engineering, Anhui Jianzhu University, Hefei 230601, China

Received:2026-01-16 Published:2026-11-06

摘要/Abstract

摘要： [目的] 针对传统方法对侧扫声呐数据进行海底检测时存在精度不足，易导致图像拼接困难、形状畸变等难题，本文旨在实现高精度、稳定的侧扫声呐实时海底检测。[方法] 为此，提出了一种基于U-Net架构的双通道一维实时海底检测模型 (D1D-UNet)。该模型以侧扫声呐左右舷的双通道序列数据作为输入，增强了模型对侧扫声呐数据的全局与局部特征的感知能力，从而显著提高了海底检测精度。[结果] 在湄洲湾和鲅鱼圈实测数据集上的试验结果表明，模型准确率最高分别达到97.26%和96.16%。与现有主流方法相比，在湄洲湾3个测试集上，检测准确率最高分别提升了5.35、6.58、12.05个百分点。[结论] 本文研究为侧扫声呐的实时海底检测提供了一种有效的解决方案，进一步拓展了侧扫声呐应用范围。

关键词: 侧扫声呐, 海底检测, 深度学习

Abstract: [Purposes] To address the challenges of low accuracy in traditional bottom detection methods using side scan sonar data,which leads to difficulties such as image stitching problems and shape distortion,this study aims to achieve high-accuracy and stable real-time bottom detection for side scan sonars. [Methods] Accordingly,a dual-channel one-dimensional real-time bottom detection model based on the U-Net architecture (D1D-UNet)is proposed.The model takes dual-channel sequential data from the port and starboard sides of the side-scan sonar as input,enhancing its ability to capture both global and local features of the sonar data,thereby significantly improving bottom detection accuracy. [Findings] Experiments conducted on measured datasets from Meizhou Bay and Bayuquan demonstrate that the model achieved maximum accuracies of 97.26% and 96.16%,respectively.Compared with existing mainstream methods,the detection accuracy on the three test sets of Meizhou Bay improved by 5.35,6.58,and 12.05 percentage point. [Conclusions] This study provides an effective solution for real-time bottom detection using side scan sonars and further expands the application scope of side-scan sonars．

Key words: side scan sonar, bottom detection, deep learning

中图分类号:

P229

严俊, 李洪, 孟俊霞, 陈书祥. D1D-UNet：侧扫声呐实时海底检测的双通道一维U-Net模型[J]. 测绘通报, 2026, 0(10): 112-118,186.

Yan Jun, Li Hong, Meng Junxia, Chen Shuxiang. D1D-UNet: a dual-channel one-dimensional U-Net model for real-time bottom detection with side scan sonar[J]. Bulletin of Surveying and Mapping, 2026, 0(10): 112-118,186.

导出引用管理器 EndNote|Reference Manager|ProCite|BibTeX|RefWorks

链接本文: http://tb.chinasmp.com/CN/10.13474/j.cnki.11-2246.2026.1017

http://tb.chinasmp.com/CN/Y2026/V0/I10/112

参考文献

[1] Terracciano D S,Bazzarello L,Caiti A,et al.Marine robots for underwater surveillance[J].Current Robotics Reports,2020,1(4):159-167.
[2] Chutia S,Kakoty N M,Deka D.A review of underwater robotics,navigation,sensing techniques and applications[C]//Proceedings of 2017 Advances in Robotics.New Delhi: ACM Press,2017．
[3] Ikpe E J.Advancements in subsea technologies for deepwater exploration for optimum and efficient production； a comprehensive review[C]//Proceedings of 2024 SPE Nigeria Annual International Conference and Exhibition.Lagos:SPE,2024.
[4] Liu Yang,Li Sanzhong,Zou Zhuoyan,et al.Techniques and methods for seafloor topography mapping:past,present,and future[J].Intelligent Marine Technology and Systems,2025,3(1):8.
[5] 王晓,王爱学,蒋廷臣,等.侧扫声呐图像应用领域综述[J].测绘通报,2019(1):1-4．
[6] 赵玉新,赵廷.海底声呐图像智能底质分类技术研究综述[J].智能系统学报,2020,15(3):587-600．
[7] Einsidler D,Dhanak M,Beaujean P P.A deep learning approach to target recognition in side-scan sonar imagery[C]//Proceedings of 2019 OCEANS MTS/IEEE Charleston.Charleston:IEEE,2019.
[8] Chavez P S,Isbrecht J,Galanis P,et al.Processing,mosaicking and management of the Monterey Bay digital sidescan-sonar images[J].Marine Geology,2002,181(1/2/3):305-315.
[9] 赵建虎,王晓,张红梅.侧扫声呐图像海底线自动提取方法研究[J].武汉大学学报(信息科学版),2017,42(12):1797-1803．
[10] 王爱学,金绍华,刘天阳,等.采用方向自适应密度聚类自动检测侧扫声呐图像海底线[J].武汉大学学报(信息科学版),2025,50(4):674-683．
[11] 张济博,潘国富,丁维凤.基于LOG算子的侧扫声呐海底线检测[J].海洋通报,2010,29(3):324-328．
[12] 库安邦,周兴华,王方旗,等.一种联合单波束测深的侧扫声呐海底线提取方法[J].海洋通报,2019,38(1):47-52．
[13] 罗逸豪,刘奇佩,张吟,等.基于深度学习的水下图像目标检测综述[J].电子与信息学报,2023,45(10):3468-3482．
[14] Guo Tiantian,Song Yunze,Kong Zejian,et al.Underwater target detection and localization with feature map and CNN-based classification[C]//Proceedings of 2022 International Conference on Advances in Computer Technology,Information Science and Communications (CTISC).Suzhou:IEEE,2022．
[15] Zeng Lingcai,Sun Bing,Zhu Daqi.Underwater target detection based on Faster R-CNN and adversarial occlusion network[J].Engineering Applications of Artificial Intelligence,2021,100:104190．
[16] Kumar S,Chinthaginjala R,Dhanamjayulu C,et al.Enhancing underwater target localization through proximity-driven recurrent neural networks[J].Heliyon,2024,10(7):e28725.
[17] Bajpai V,Sharma A,Subudhi B N,et al.Underwater U-Net:deep learning with U-Net for visual underwater moving object detection[C]//Proceedings of 2021 OCEANS:San Diego - Porto.San Diego:IEEE,2021.
[18] 陈禹蒲,马晓川,李璇.基于YOLOv3锚框优化的侧扫声呐图像目标检测[J].信号处理,2022,38(11):2359-2371．
[19] Sun Yushan,Zheng Haotian,Zhang Guocheng,et al.CGF-unet:semantic segmentation of sidescan sonar based on unet combined with global features[J].IEEE Journal of Oceanic Engineering,2024,49(3):963-975．
[20] Yan Jun,Meng Junxia,Zhao Jianhu.Bottom detection from backscatter data of conventional side scan sonars through 1D-UNet[J].Remote Sensing,2021,13(5):1024．
[21] Tang Yulin,Wang Liming,Jin Shaohua,et al.AUV-based side-scan sonar real-time method for underwater-target detection[J].Journal of Marine Science and Engineering,2023,11(4):690.

D1D-UNet：侧扫声呐实时海底检测的双通道一维U-Net模型

D1D-UNet: a dual-channel one-dimensional U-Net model for real-time bottom detection with side scan sonar

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	王轶虹, 魏勇, 王冬梅, 张鹏. MFG-SCDNet：多尺度感知与变化引导融合的水体语义变化检测网络[J]. 测绘通报, 2026, 0(9): 75-82.
[2]	黄铭琴, 安佳乐, 郑江华, 尼格拉·吐尔逊, 刘丽, 王南, 林峻, 刘宣, 徐亚鹏, 李璇. 大尺寸无人机影像黄兔尾鼠洞穴识别方法[J]. 测绘通报, 2026, 0(9): 108-114.
[3]	张用川, 高洁, 张治清, 周正强, 王陆潇, 官冬杰. 多模态融合框架下的城市功能区识别[J]. 测绘通报, 2026, 0(8): 96-102.
[4]	龙飞, 王文辉, 石一凡, 林思群, 刘文壮, 邵林. 一种融合全局与局部语境信息的水体分割算法[J]. 测绘通报, 2026, 0(8): 130-136.
[5]	周洪, 张国和, 游思琪, 王冉, 丁鹏飞. 双主干深度学习网络支撑下的ZY1E影像耕地智能提取[J]. 测绘通报, 2026, 0(8): 137-144.
[6]	黄彬, 石国杰, 杜力, 王林柯, 韩沐阳, 彭鸿伟, 李克凡, 李佳玺, 王晋国, 侯兆阳. 基于YOLO11改进网络的探地雷达地下病害智能识别[J]. 测绘通报, 2026, 0(8): 154-160,173.
[7]	刘伟丽, 田志勇, 罗勇强. 基于深度学习的武汉市屋顶光伏地理潜力评估[J]. 测绘通报, 2026, 0(7): 167-172.
[8]	杜韵琦, 施泓羽, 张红娟, 李必军. RefDC:聚焦远距离深度特征细化的深度补全方法[J]. 测绘通报, 2026, 0(6): 67-73.
[9]	徐超. 基于质量图引导的深度学习相位解缠方法[J]. 测绘通报, 2026, 0(6): 92-97.
[10]	杨刚, 王淼, 陈思, 周佺, 李江川. 基于改进的YOLO11的天际线检测算法[J]. 测绘通报, 2026, 0(5): 12-16.
[11]	陈冲, 杨扬. 基于视觉状态空间模型的城市遥感影像语义分割方法[J]. 测绘通报, 2026, 0(5): 103-109.
[12]	郭明, 张潇澜, 仇宫润, 郭帅, 朱丽. 基于激光雷达点云的地下空间目标智能识别[J]. 测绘通报, 2026, 0(4): 65-72.
[13]	梁卫涛, 王跃东, 谢亚奇, 丁武, 杨红磊, 吴波, 薛启浪. 基于InSAR和深度学习的韩城市地表形变监测与预测[J]. 测绘通报, 2026, 0(4): 112-118,139.
[14]	李宏强, 杨魁, 车国伟. 融合注意力机制的加权ResNet在SAR洪水淹没区提取中的应用[J]. 测绘通报, 2026, 0(4): 134-139.
[15]	杨宇挺, 胡婷, 潘子永, 童旭东, 马爱龙. 基于多源遥感地理空间数据的建筑物用途识别[J]. 测绘通报, 2026, 0(3): 44-50.