养殖鱼类摄食行为的特征提取研究与应用进展  

Advances in the study and application of feature extraction in feeding behavior of cultured fish

作　　者：曹晓慧 刘晃[2] CAO Xiaohui;LIU Huang(College of Fisheries and Life Science,Shanghai Ocean University,Shanghai 20l306,China;Fishery Machinery and Instrument Research Institute,Chinese Academy of Fishery Sciences,Shanghai 200092,China)

机构地区：[1]上海海洋大学水产与生命学院,上海201306 [2]中国水产科学研究院渔业机械仪器研究所,上海200092

出　　处：《渔业现代化》2021年第2期1-8,共8页Fishery Modernization

基　　金：国家重点研发计划“工厂化养殖水质智能调控技术与装备研发(2019YFD0900501)”。

摘　　要：为解决养殖过程中饲料浪费和水质污染的问题,鱼类摄食行为的研究一直受到广大研究者的重视。量化鱼类摄食行为,提取可用摄食参数,构建精确的摄食模型是优化投喂策略的有效途径。总结了摄食过程中主要觅食感觉器官、捕食量化指标和摄入声信号特征。综述了养殖期间影响摄食行为的环境因素。对比分析了两种监测技术的研究进展及优缺点,即基于可见光与近红外光的计算机视觉技术和基于声学监测与声呐成像的声学技术,针对存在的问题提出了能够提高摄食模型准确性和普遍性的方法,以及被动声学的研究潜力及研究方向,以期为投饲系统向智能化的迈进提供参考。In order to solve the problem of feed waste and water pollution in the process of aquaculture,the study of fish feeding behavior has been paid much attention by the majority of researchers.Quantifying fish feeding behavior,extracting available parameters,and constructing a precise model are the effective ways to optimize feeding strategy.This paper summarized the main foraging sensory organs,quantitative indexes of predation and the characteristics of acoustic signals during feeding,and summarized the environmental factors that affect feeding behavior in the period of cultivation.The research progress,advantages and disadvantages of the two monitoring technologies were compared and analyzed,including computer vision based on visible and near-infrared light,and acoustic technology based on acoustic monitoring and sonar imaging.In view of the existing problems,this paper proposed methods to improve the feeding model accuracy and universality,and research potential and direction of passive acoustics,so as to provide reference for the intelligent feeding system.

关 键 词：鱼类摄食行为 摄食模型 计算机视觉 声学 

分 类 号：S951.2[农业科学—水产养殖]