基于机器视觉的工厂化鱼类养殖投料控制系统研制  

Development of a Feeding Control System for Industrialized Fish Farming Based on Machine Vision

作　　者：魏巍宏 康伟达 WEI Wei-hong;KANG Wei-da(Xiamen Key Laboratory of Intelligent Fishery,Xiamen Ocean Vocational College,Xiamen 361100,China)

机构地区：[1]厦门海洋职业技术学院厦门市智慧渔业重点实验室,福建厦门361100

出　　处：《唐山师范学院学报》2024年第6期66-70,共5页Journal of Tangshan Normal University

基　　金：福建省中青年教师教育科研项目(JAT210811);厦门市智慧渔业重点实验室2024年度主任基金项目(XMKLIFZR-202401)。

摘　　要：为了实现工厂化循环水养殖中的鱼群数量估算和投料自动控制,以加州鲈鱼为研究对象,基于YOLOv8构建了一种鱼群数量估算模型,并结合投料设备研制了一套鱼类养殖投料控制系统。该系统首先通过DCP算子增强鱼群图像边界;然后使用YOLOv8深度学习模型识别鱼体轮廓并对鱼群数量进行分级,结合养殖户经验为各数量等级鱼群设定相应饲料投放量;最终系统通过PLC控制投料机投放给定重量饲料,实现饲料投喂的全自动化管理。测试证实:该系统能够实时估算鱼群数量,控制投料机完成饲料投放量合理的自动投喂工作。To achieve the estimation of fish population quantity and automatic feeding control in industrialized recirculating aquaculture,we focused on largemouth bass as the research subject.Based on YOLOv8,we developed a fish population estimation model and designed an automatic feeding control sys-tem integrated with feeding equipment.The system first enhanced the boundaries of the fish image using DCP operators,then employed the YOLOv8 deep learning model to recognize fish contours and classify the fish population.Combining the farmers'expertise,we set specific feed quantities for each fish population level.Finally,the system controled the feeder via PLC to dispense the predetermined amount of feed,a-chieving fully automated management of feed dispensing.Testing has confirmed that this system can accu-rately estimate the fish population quantity in real-time and automatically dispense the appropriate amount of feed through the feeding equipment.

关 键 词：鱼群数量估算 机器视觉 YOLOv8 自动投料 

分 类 号：S969.31[农业科学—水产养殖]