黄  健¹，关长涛²，汤涛林¹，李  娇²
（1 十大网投正规信誉官网，农业部渔业装备与工程重点开放实验室，上海 200092；
2 中国水产科学研究院黄海水产研究所，青岛 266071）

摘要：构造了基于声纳信号处理的深水网箱生物量评估数学模型。运用单回波脉宽检测和跟踪方法，结合二维低通滤波器，对声纳数据进行处理，去除背景噪声和其他回波，在单体鱼连续的回波踪迹中可以检测出鱼的游速和方向。同时，结合图像分析有效压缩噪声，提取有用特征来识别检测到的鱼迹，形成一种自动计数方法，提高生物量的检测概率。该方法通过提取和使用原始的回声图像，检测出运动目标，同时压缩回声图中的噪声，从而在低信噪比数据中实现对网箱中生物量的检测和估计，适合实际工程应用。?

关键词：单回波检测；低通滤波；深水网箱；生物量评估

Research on a novel biomass estimation model for deep water net cages based on single echo detection and tracking
           HUANG Jian^{1, 2}, TANG Tao-lin^{1, 2}, NI Han-hua^{1, 2}, LI Jiao³ 
( 1 Fishery Machinery Instrument Research Institute, Chinese Academy of Fishery Sciences,
Key Laboratory of Fishing Equipment and Engineering, Ministry of Agriculture, Shanghai 200092, China;
 2 Yellow Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Qingdao 266071,China )

Abstract: The mathematical model for biomass estimation in deep water net cages is constructed based on sonar signal processing. Signal echo detection and tracking method for pulse width together with 2-D low-pass filter is applied to process sonar data in order to dismiss background noises and other echoes. The swimming speed and direction of a single fish can be detected from its continuous echo track. Image analysis is also added to compress noises efficiently and pick up useful features to identify detected fish tracks, and then produce an automatic counting method to improve the biomass detection probability. This method is used to detect moving objects and compress image noises by picking up and employing original echo images, and accordingly achieves the detection and estimation for the biomass by means of low signal to noise ratio data, which is suitable for engineering application.

Key words: single echo detection; low-pass filter; deep water net cages; biomass estimation

基金项目：国家“863”高技术研发计划项目（2006AA100302）

作者简介：黄健（1982－），男，工学硕士，研究方向：水声电子技术，信号处理和水质智能监控。E-mail:huang_jian1982@126.com

                                          （《渔业现代化》 2008，35（6），9-13）