针对水下地形匹配定位问题，提出一种基于多粒度声呐图像匹配的定位方法。构建了U-Net架构的图像特征抽取模型，以无监督方式抽取水下声呐图像的高层特征。设计了一种具有时空约束的多粒度匹配算法，在特征空间内首先进行粗粒度匹配，并结合时空约束约简搜索空间，然后进行细粒度精确匹配，根据匹配结果实现水下定位。通过建模水下典型地形并进行模拟声呐探测构建了数据集，包括水下地形高程数据和对应的声呐探测图像。在自构数据集上的实验结果表明，所提方法的定位精度可达0.679 m，平均单次定位时长小于0.5 s，性能优于基线算法。
Aiming at the problem of underwater terrain matching, a location method based on multi-granularity sonar image matching is proposed. An image feature extraction model based on U-Net architecture is constructed to extract the high-level features of underwater sonar images in an unsupervised manner. A multi-grain matching algorithm with spatio-temporal constraints is designed. Coarse-grained matching was first performed in the feature space, and the search space is reduced with spatio-temporal constraints. Then fine-grained precise matching is performed, and underwater positioning is realized according to the matching results. The data set is constructed by modeling typical underwater terrain and simulating sonar detection, including underwater terrain elevation data and corresponding sonar detection images. The experimental results on the autogenous data set show that the proposed method can achieve a positioning accuracy of 0.679 m, the average single positioning time is less than 0.5 seconds, which is superior to the baseline algorithm.
2025,47(14): 95-102 收稿日期：2024-9-11
DOI：10.3404/j.issn.1672-7649.2025.14.015
分类号：TP391.41
基金项目：中国博士后科学基金资助项目（2020M682348）；海洋防务技术创新中心创新基金资助项目（JJ-2022-709-01）
作者简介：王可(1985-),男,博士,副教授,研究方向为计算智能、人机融合智能、神经计算与优化算法
参考文献：
[1] SALAVASIDIS G , MUNAFÒ, ANDREA, HARRIS C A , et al. Terrain‐aided navigation for long‐endurance and deep‐rated autonomous underwater vehicles[J]. Journal of Field Robotics, 2018.
[2] MAI N T , JI Y , WOO H , et al. Acoustic Image Simulator Based on Active Sonar Model in Underwater Environment[C]// 2018.
[3] L PAULL, S SAEEDI, M SETO, et al. AUV Navigation and Localization: A Review[J]. IEEE Journal of Oceanic Engineering, 2014, 39(1): 131-149.
[4] 桑恩方, 庞永杰, 卞红雨. 水下机器人技术[J]. 机器人技术与应用, 2003(3): 8-13.
SANG E F, PANG Y J, BIAN H Y. Underwater robot technology[J]. Robotics and applications, 2003(3): 8-13.
[5] 陈小龙. AUV水下地形匹配辅助导航技术研究[D]. 哈尔滨: 哈尔滨工程大学, 2012.
[6] RONNEBERGER O, FISCHER P, BROX T. U-net: Convolutional Networks for Biomedical Image Segmentation[C]// Medical Image Computing and Computer-Assisted Intervention– MICCAI 2015. Springer, Cham, 2015.
[7] INGEMAR N, MEMBER. Terrain navigation for underwater vehicles using the correlator method[J]. IEEE Journal of Oceanic Engineering, 2004, 29(3): 906-915.
[8] HUANG Y, ZHANG Y, ZHAO Y. Review of autonomous undersea vehicle navigation methods[J]. Journal of Unmanned Undersea Systems, 2019, 27(3): 232-253.
[9] 宋子奇. 基于声呐图像处理的水下地形地貌匹配定位方法研究[D]. 哈尔滨: 哈尔滨工程大学, 2017.
[10] GOLDEN J P . Terrain Contour Matching (TERCOM): A Cruise Missile Guidance Aid[C]// Technical Symposium, 1980.
[11] HAN Y R, WANG B, DENG Z H, et al. An Improved TERCOM-Based Algorithm for Gravity-Aided Navigation[J]. IEEE Sensors Journal, 2016, 16(8): 2537-2544.
[12] LI P J, ZHANG X F, XU X S. Novel terrain integrated navigation system using neural network aided Kalman filter[C]// 2010 Sixth International Conference on Natural Computation. Yantai, 2010.
[13] HUANG L Y, HE B, ZHANG T. An autonomous navigation algorithm for underwater vehicles based on inertial measurement units and sonar[C]// 2010 2nd International Asia Conference on Informatics in Control, Automation and Robotics (CAR 2010). Wuhan, 2010.
[14] TAN S L, DU K Y, HUANG W D, et al. Segmented Underwater Terrain Matching Based on Q-Learning[C]// IEEE International Conference on Signal Processing, Communications and Computing (ICSPCC). Xi'an, 2022.
[15] YANG W L, FAN S S, XU S X, et al. Autonomous Underwater Vehicle Navigation Using Sonar Image Matching based on Convolutional Neural Network[J]. IFAC-PapersOnline, 2019, 52(21): 156-162.
[16] SANTOS M M D, GIACOMO G G D, DREWS P L J, et al. Matching Color Aerial Images and Underwater Sonar Images Using Deep Learning for Underwater Localization[J]. IEEE Robotics and Automation Letters, 2020, 5(4): 6365-6370.
[17] MATHIEU M, COUPRIE C, LECUN Y. Deep multi-scale video prediction beyond mean square error[J]. Computer Science, 2015.
[18] HASAN M, CHOI J, NEUMANN J, et al. Learning Temporal Regularity in Video Sequences[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, 2016.
[19] 梁家菲, 李婷, 杨佳琪, 等. 融合自注意力和自编码器的视频异常检测[J]. 中国图象图形学报, 2023, 28(4): 1029-1040
LIANG J F, LI T, YANG J Q, et al. Video anomaly detection combining autoattentional and autoencoder[J]. Chinese Journal of Image and Graphics, 2023, 28(4): 1029-1040.
[20] RÔMULO CERQUEIRA, TROCOLI T, NEVES G, et al. A novel GPU-based sonar simulator for real-time applications[J]. Computers & Graphics, 2017, 68: 66-76.
[21] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[J]. Computer Science, 2014.
[22] HE K, ZHANG X, REN S, et al. Deep Residual Learning for Image Recognition[C]// IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, 2016.
[23] HUANG G, LIU Z, LAURENS V, et al. Densely Connected Convolutional Networks[C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, 2016.
