声纳图像边缘检测方法对比研究  

A comparison study of edge detection algorithms for sonar image

作　　者：李荣 李响 吉玲 Li Rong;Li Xiang;Ji Ling(Bengbu Navy Petty Officer Academy,Anhui Bengbu,233012;College of Internet of ThingsEngineering,Hohai university,Changzhou Nanjing,213022)

机构地区：[1]海军士官学校,安徽蚌埠233012 [2]河海大学物联网工程学院,江苏常州213022

出　　处：《电子测试》2020年第11期60-61,66,共3页Electronic Test

摘　　要：猎雷声纳图像是探测识别水雷的重要依据,对猎雷声纳图像进行图像边缘检测处理有助于水雷目标的判别。分析了经典的边缘检测法、相位一致性法以及分数阶微分法进行图像边缘检测的原理和方法。通过实验,对比了几种方法在无噪声和有噪声声纳图像中的处理效果及各自的特点,并将其应用于猎雷声纳图像的处理。结果表明:分数阶傅立叶变换可以有效的用于猎雷声纳图像中目标特征的提取。Mine-hunting sonar image can be used for mine detection and recognition.Image edge detection is a useful method for the sonar image processing.The principles and work processes of the classic edge detection methods,the phase congruency method and the fractional differentiation method are analyzed.These methods are tested on noise-free and noise sonar images and also used in sonar image processing.Their advantages and disadvantages are discussed and indicated that fractional differentiation is adaption to sonar image processing.

关 键 词：边缘检测 声纳图像 CANNY算子 相位一致性 分数阶微分 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TJ617[自动化与计算机技术—计算机科学与技术]