提出大噪声环境下前视声呐图像目标识别的研究方法，针对水下无人航行器（UUV）在近岸浅水区航行中由于前视声呐图像噪声较大，难以准确识别目标的问题，通过改进的中值滤波和Otus阈值检测算法，对前视声呐图像进行滤波和二值化。利用区域增长算法分割疑似目标区域图像，分别提取分割图像的长度、形状、方向、灰度均值和灰度能量中值等参数，利用支持向量机（SVM）的方法对这些图像参数训练和识别，结果表明该方法能够有效识别大噪声环境下的前视声呐图像目标。
This paper proposes a research method for the target recognition of forward-looking sonar images in a large noise environment. Aiming at the problem that the underwater unmanned vehicle(UUV) is difficult to accurately identify the target due to the large noise of the forward-looking sonar image when the UUV sails in shallow waters near the shore. Through the improved median filter and Otus threshold detection algorithm, the forward-looking sonar image is filtered and binarized. The region growth algorithm is used to segment the image of the suspected target area, and the length, shape, direction, gray average value and gray energy segmentation value of the segmented iamge are separately extracted, and the support vector machine(SVM) method is used to train and recognize these image parameters. The results show that the method can effectively identify the forward-looking sonar image target in a large noise environment.
2022,44(1): 125-130 收稿日期：2021-08-29
DOI：10.3404/j.issn.1672-7649.2022.01.024
分类号：U666.7
基金项目：河南省高等学校重点科研项目(22A413003)
作者简介：王凯(1992-)，男，助理工程师，主要从事水下航行器声呐图像处理研究
参考文献：
[1] 徐会希. 自主水下机器人[M]. 北京: 科学出版社, 2019.
[2] 宋三明. 基于随机场的水下声呐图像处理[M]. 北京: 科学出版社, 2020.
[3] JIANG Wen-long, LI Guang-lin, LUO Wei-bing. Application of improved median filtering algorithm to image de-noising[J]. Advanced Materials Research, 2014, 998: 838–841
[4] LI W, NI R, LI X, et al. Robust median filtering detection based on the difference of frequency residuals[J]. Multimedia Tools and Applications, 2018, 78(4): 8363–8381
[5] CERVENKA P, DEMOUSTIER C. Sided-scan sonar image processing techniques[J]. IEEE Journal of Oceanic Engineering, 1993, 18(2): 108–122P
[6] 马珊. 水下机器人前视声呐多目标跟踪技术研究[D]. 哈尔滨: 哈尔滨工程大学, 2016.
[7] 石洋, 胡长青. 基于粒子群最小二乘支持向量机的前视声呐目标识别[J]. 声学技术, 2018, 37(2): 122–128
Shi Yang, Hu Changqing. Forward looking sonar target recognition based on particle swarm least squares support vector machine[J]. Acoustics Technology, 2018, 37(2): 122–128
[8] CHANG Chih-Chung, LIN Chik-Jen, LIBSVM: A library for support vector machines[J]. ACM Transactions on Intelligent Systems and Technology, 2011, 2(3): 27.
