基于双频识别声呐的鱼类目标识别与计数  

Fish target recognition and counting based on Dual-frequency Identification Sonar

作　　者：沈蔚[1,2] 朱振宏 张进[1] 曹正良[1] 彭战飞 SHEN Wei;ZHU Zhenhong;ZHANG Jin;CAO Zhengliang;PENG Zhanfei(College of Marine Sciences,Shanghai Ocean University,Shanghai Engineering Research Center of Estuarine and Oceanographic Mapping,Shanghai 201306,China;Engineering Technology Research Center of Marine Ranching,Shanghai Ocean University,Shanghai 201306,China)

机构地区：[1]上海海洋大学海洋科学学院,上海河口海洋测绘工程技术研究中心,上海201306 [2]上海海洋大学海洋牧场工程技术研究中心,上海201306

出　　处：《渔业现代化》2020年第6期81-87,共7页Fishery Modernization

基　　金：东海渔业资源评价和增殖养护技术研究与示范科研专项(201303047);2020年上海海洋大学科技发展专项基金。

摘　　要：鱼类资源调查是渔业管理的基础性工作,利用水声探测技术进行鱼类资源调查日益成为主流手段。针对双频识别声呐(DIDSON)在水库中采集到的声学数据,利用Echoview声学数据后处理软件及其相关算法进行鱼类识别和计数研究。首先利用Kovesi图像去噪法和背景差分法对声学图像中斑点噪声和水体背景进行去除,然后根据鱼类目标回声阈值进行目标识别,利用α-β轨迹追踪算法进行运动目标追踪以防止重复计数,最后进行目标计数和体长信息的提取。结果显示,与借助声学图像的人工计数相比,本方法鱼类目标计数误差在10%以内,平均计数误差为7.2%,具有相当高的统计精度。研究表明,DIDSON可以用于鱼类识别与计数,在浅水水域的鱼类资源探测与管理方面具有非常广阔的应用空间。Fish resource survey is a basic work of fishery management,and the investigation of fish resources using underwater acoustic detection technology has become a mainstream method.Aiming at the acoustic data collected by Dual-frequency Identification Sonar(DIDSON)in the reservoir,the acoustic data post-processing software Echoview and related algorithms were used to conduct fish identification and counting research.First,the Kovesi image denoising method and the background difference method were utilized to remove speckle noise and water background in acoustic images,then targets were identified according to the echo threshold,theα-βtrajectory tracking algorithm was used to track moving targets to prevent double counting,and finally fish counting and length extraction were performed.The results show that,compared with munual counting,fish counting error of this method is within 10%,and the average error is 7.2%,with a very high statistical accuracy.Studies have proven that DIDSON can be used for fish identification and counting,and has a very broad application space for fish resource detection and management in shallow waters.

关 键 词：双频识别声呐 声呐图像 目标识别 鱼类计数 资源评估 

分 类 号：S932.4[农业科学—渔业资源]