基于图像处理的回声定位信号检测方法  

Detection Method for Echolocation Clicks Based on Image Processing

作　　者：段德鑫 姜莹 刘宗伟 杨春梅 吕连港 DUAN De-xin;JIANG Ying;LIU Zong-wei;YANG Chun-mei;LÜLian-gang(First Institute of Oceanography, MNR, Qingdao 266061, China;Key Laboratory of Marine Science and Numerical Modeling, MNR, Qingdao 266061, China;Shandong Key Laboratory of Marine Science and Numerical Modeling, Qingdao 266061, China;Laboratory for Regional Oceanography and Numerical Modeling,Pilot National Laboratory for Marine Science and Technology(Qingdao),Qingdao 266237,China)

机构地区：[1]自然资源部第一海洋研究所,山东青岛266061 [2]自然资源部海洋环境科学与数值模拟重点实验室,山东青岛266061 [3]山东省海洋环境科学与数值模拟重点实验室,山东青岛266061 [4]青岛海洋科学与技术试点国家实验室区域海洋动力学与数值模拟功能实验室,山东青岛266237

出　　处：《海洋科学进展》2022年第1期145-153,共9页Advances in Marine Science

基　　金：国家自然科学基金项目——南极普里兹湾和麦克默多湾虎鲸声信号特征研究(41906170)。

摘　　要：鲸豚类海洋哺乳动物发出的回声定位信号是一种频率较高的短时脉冲信号,研究回声定位信号检测方法有助于快速、准确地检测海洋哺乳动物发声,进而更好地保护海洋哺乳动物。本文提出了一种基于图像处理的鲸豚类动物回声定位信号检测方法。该方法首先对收集到的数据进行分帧,计算每帧信号的时频图;使用Frangi滤波器对时频图进行滤波,以降低背景噪声并凸显线条状信号;然后根据鲸豚类海洋哺乳动物回声定位信号在时频图中类似直线的特点,使用直线检测获得时频图像中检出线条状信号的精确位置,进而提取出其图像特征;最后根据信号特征训练随机森林分类器生成检测模型以检测区分出回声定位信号。使用柏氏中喙鲸、领航鲸、露脊鲸和伊洛瓦底海豚四种动物的声数据进行了实验,结果表明,与传统能量算子方法相比,本文提出的方法在低信噪比条件下有着更高的召回率且具有较高的准确率,可取得更好的检测效果。本文方法能够为鲸豚类海洋哺乳动物的声学研究提供一定的技术支持。Odontocetes echolocation click is a kind of short-time signal with high frequency.Research on the detection method of echolocation signal is helpful for the rapid and accurate detection of marine mammal vocalization,so as to better protection of marine mammals.A method based on image processing is proposed to detect odontocetes echolocation clicks.The collected data are divided into fixed-length frames and the time-frequency diagram of each frame is calculated.The spectrogram images of click signals are filtered to remove noise and enhance the line-shape signals.Considering that the echolocation signals are like lines in time-frequency domain,line detection is subsequently used to obtain the precise position of the signal,and then to acquire the feature of signals.Finally,a Random Forest Classifier was trained to obtain a detector to distinguish between click and non-click clips.Experiments were carried out using acoustic data of Mesoplodon densirostris,Globicephalamelas,Eubalaena japonica and Orcaella brevirostris.The performance of the proposed method was evaluated using real audio recordings.The experimental results indicate that comparing with the traditional energy detector method,the proposed algorithm shows higher recall and precise under low signal-to-noise ratio(SNR).The proposed method can provide technical support for marine bioacoustics study.

关 键 词：图像处理 Frangi滤波 直线检测 随机森林分类器 回声定位信号 

分 类 号：Q62[生物学—生物物理学]