渔业现代化

渔业现代化 ›› 2023, Vol. 50 ›› Issue (4): 107-117. doi: 10.3969/j.issn.1007-9580.2023.04.013

• • 上一篇    下一篇

一种基于前视声呐的养殖网箱内鱼类数量评估方法 

  1. （1 浙江海洋大学船舶与海运学院, 浙江 舟山 316022；
    2  海卓科（浙江自贸区）科技有限公司, 浙江 舟山 316020；
    3 浙江海洋大学海洋科学与技术学院, 浙江 舟山 316022；
    4 浙江海洋大学国家海洋设施养殖工程技术研究中心, 浙江 舟山 316022；
    5  浙江海洋大学水产学院, 浙江 舟山 316022）
  • 出版日期:2023-08-20 发布日期:2023-09-04
  • 作者简介:崔智强(1998—)，男，硕士研究生，研究方向：水声信号处理与声呐图像处理。 E-mail：m17769825232@163.com
  • 基金资助:
    浙江省重点研发计划项目(2020C02004);  东海实验室开放基金项目(DH-2022KF01018);  浙江省大学生科技创新计划(新苗人才计划)(2022R411C050，2022R411C052) 

A method for assessing the number of fish in aquaculture cages based on forward-looking sonar#br#
#br#

  1. （1School of Naval Architecture and Maritime, Zhejiang Ocean University, Zhoushan 316022,Zhejiang, China;
    2 Hydroacoustics Technology Co., Ltd., Zhoushan 316022, Zhejiang, China;
    3 Marine Science and Technology College, Zhejiang Ocean University, Zhoushan 316022, Zhejiang, China;
    4 National Engineering Research Center for Marine Aquaculture, Zhejiang Ocean University, Zhoushan 316022, Zhejiang, China
    5 School of Fishery, Zhejiang Ocean University, Zhoushan 316022, Zhejiang, China）

  • Online:2023-08-20 Published:2023-09-04

PDF

摘要： 为准确、高效评估养殖网箱场内鱼类数量，科学指导网箱养殖，提出了一种基于前视声呐的养殖网箱内鱼类数量统计方法。方法应用时，以声呐图像处理技术为基础，首先以目标网箱内养殖鱼类在消声水池实验室的声学标定结果为依据，对前视声呐测量声图进行单个鱼体的声学图像阈值划分；其次根据云台控制声呐旋转，对围栏区域进行遍历探测，随后将声学标定结果确定的阈值作为划分依据，将不同角度所测声学图像中鱼类进行划分、统计，估算得到整个养殖网箱内的鱼类数量；最后结合探测角度以及位置信息，得到网箱养殖区域内的鱼类空间分布特征，用于指导渔民进行后续的网箱养殖。此方法在舟山青浜岛大黄鱼养殖网箱中得以应用。结果显示：该方法估计得到的养殖网箱内大黄鱼数量与人工评估结果基本一致，经多次计算，平均相对误差仅为2.45%，且实现过程更加高效、便捷。本研究能够有效为养殖网箱提供鱼类数量信息，为亟待需要进行鱼类数量监管的网箱养殖产业提供科学、可靠的数量统计方法，具有广泛的应用场景。


关键词: 水产养殖, 统计评估, 空间分布, 前视声呐

Abstract: In order to accurately and efficiently evaluate the number of fish in the cage and scientifically guide the cage culture, a method of counting the number of fish in the cage based on forward-looking sonar was proposed. Methods In the application of the method, based on sonar image processing technology, the acoustic image threshold of a single fish body was divided based on the acoustic calibration results of the fish cultured in the target cage in the  anechoic pool laboratory. Secondly, controling sonar rotation according to the PTZ, the traversal detection of the fence area is carried out. Then, the threshold value determined by the acoustic calibration results was used as the basis for classification, and the fish quantity in the whole aquaculture cage was estimated by dividing and counting the fish in the acoustic images measured from different angles.. Finally, combined with the detection Angle and location information, the spatial distribution characteristics of fish in the cage culture area were obtained, which could be used to guide fishermen to carry out follow-up cage culture. The method was applied in the cage culture of large yellow croaker in Zhoushan Qingbang Island, and the results showed that the estimated quantity of large yellow croaker in the cage was basically consistent with the artificial evaluation results. After several calculations, the average relative error is only 2.45%. The realization process is more efficient and convenient, and it has a wide application prospect in the subsequent cage culture guidance. It can effectively provide fish quantity information for aquaculture cages, and provide scientific and reliable quantity statistics methods for the cage culture industry, which is in urgent need of fish quantity supervision, and has a wide range of application scenarios.


Key words: aquaculture, statistical assessment, spatial distribution, forward looking sonar