改进区域生长算法的海洋航道浅剖底质层界智能识别  

Intelligent identification of bottom layer boundaries in shallow sections of ocean waterways using improved region growth algorithm

作　　者：蒋廷臣 孟皓凡 王晓 王朝金 杨毅 宁尧尧 闫玉茹 JIANG Tingchen;MENG Haofan;WANG Xiao;WANG Chaojin;YANG Yi;NING Yaoyao;YAN Yuru(Jiangsu Ocean University School of Marine Technology and Geomatics,Lianyungang 222005,China;Jiangsu Provincial Marine Remote Sensing Engineering Research Center,Lianyungang 222005,China;Wuhan Changjiang Waterway Rescue and Salvage Bure,Wuhan 430010,China;Shanghai Da Hua Surveying&Mapping Technology Co.,Ltd.,Shanghai 201208,China;Key Laboratory of Coastal Salt Marsh Ecosystems and Resources,Ministry of Natural Resources,Nanjing 210007,China)

机构地区：[1]江苏海洋大学海洋技术与测绘学院,江苏连云港222005 [2]江苏省海洋遥感工程研究中心,江苏连云港222005 [3]武汉长江航道救助打捞局,湖北武汉430010 [4]上海达华测绘科技有限公司,上海201208 [5]自然资源部滨海盐沼湿地生态与资源重点实验室,江苏南京210007

出　　处：《测绘学报》2024年第3期493-502,共10页Acta Geodaetica et Cartographica Sinica

基　　金：国家自然科学基金(41004003);江苏省科技厅项目(BE2016701);江苏省海洋科技创新项目(JSZRHYKJ202201);江苏省水利科技项目(2020058);连云港市第“521工程”科研资助立项项目(LYG06521202131)。

摘　　要：海底勘测对海洋资源开发利用保护、海洋工程建设和国防安全等具有重要意义,浅地层剖面仪是一种能够勘测海底浅表层底质分布的声学设备,目前剖面仪的底质识别精度取决于操作者的主观性,可靠性较差,为提高效率和解译精度,需进一步研究底质层界自动识别模型。本文提出了适用于海底底质层界识别且不需要人工干预的区域生长改进算法,即在灰度映射和噪声剔除研究的基础上,研究根据迭代最大类间差算法提取图像层界骨架信息,将骨架信息作为初始生长点位,以流变特性修正生长指向,结合灰度加权映射曲线和峰谷波长约束生长邻域,通过本文算法分割层界,提取边缘、连接成线,从而实现海底底质层界识别。连云港港航道浅剖实测数据试验证明,本文算法能够有效识别底质层界,且识别精度达到厘米级,满足海底底质解译分析要求。It is of great significance to seabed exploration for the development and utilization of marine resources,marine engineering construction,and national defense security.As an acoustic device capable of surveying the distribution of bottom sediment in the shallow surface of the seabed,the accuracy of bottom sediment identification currently depends on the subjectivity of the operator,with poor reliability.In order to improve efficiency and interpretation accuracy,it is necessary to further study the intelligent identification model for the bottom layer boundary.In the paper,an improved region growth algorithm suitable for seabed bottom layer boundary recognition without human intervention is proposed.That is,based on the study of grayscale mapping and noise elimination,the skeleton information of the image layer boundary is extracted using an iterative maximum class difference algorithm,and then the skeleton information is used as an initial growth point and the growth direction is corrected using rheological properties.At the same time,the algorithm combines grayscale weighted mapping curves and peak valley wavelength constrained growth neighborhoods to segment layer boundaries,extract edges,and connect them into lines,thereby seabed bottom layer boundary recognition is ultimately achieved.The experimental results of shallow section survey data of Lianyungang port waterway show that the improved region growth algorithm can effectively identify the boundary of the bottom layer,and its recognition accuracy reaches centimeter level,which can meet the requirements of seabed sediment interpretation and analysis.

关 键 词：浅地层剖面仪 底质层界 最大类间差算法 区域生长算法 自动识别 

分 类 号：P258[天文地球—测绘科学与技术]