水产养殖智能投喂系统中单目视觉测量的研究  

Research on Monocular Vision Measurement in Intelligent Aquaculture Feeding System

作　　者：池涛[1,2] 庞云剑 沈晓晶 CHI Tao;PANG Yun-jian;SHEN Xiao-jing(College of Information Technology/Shanghai Ocean University,Shanghai 201306,China;Key Laboratory of Fisheries Information/Ministry of Agriculture,Shanghai 201306,China)

机构地区：[1]上海海洋大学信息学院,上海201306 [2]农业部渔业信息重点实验室,上海201306

出　　处：《山东农业大学学报（自然科学版）》2019年第6期980-985,共6页Journal of Shandong Agricultural University：Natural Science Edition

基　　金：国家自然科学基金(61561027);上海市自然科学基金(16ZR14100)

摘　　要：水产品投喂方式的智能化、精准化控制是提高水产养殖投喂效率的关键。本文利用计算机视觉定位精度高,不受外界环境影响(如池塘水面的波浪等障碍物的影响)等优势,提出了基于单目视觉水产养殖智能投喂系统。系统通过在池塘内的投喂设备上搭载COMS摄像头来采集食场内的目标图像,利用零阶图像矩单目深度测距算法驱使投喂设备驶向目标食场,实施对水产品的投喂操作。本文详细分析了目标图像矩与目标距离存在的关系,并实验分析了投喂设备到达食场内的目标位置与实际目标位置的精确度及误差原因,验证了本系统对水产品的精确投喂具有很高的应用价值。The intelligent and precise control of aquatic product feeding is the key to improve the efficiency of aquaculture feeding.This paper analyzes the advantages of computer vision such as high positioning accuracy and independence by the external environment(the influence of obstacles such as waves on the surface of the pond),and proposes a smart monocular-vision-based feeding system for aquaculture.The system can collect the target image in the food field by COMS camera equipped on the feeding device in the pond.In this words,the algorithm on the zero-order image moment for monocular depth ranging is implemented to drive the feeding device to the target food field to run the feeding operations.In this paper,the relationship between the target image moment and the target distance is established,and cause of error and the accuracy of the reaching the target position in the food field and the actual location are analyzed by the experiments.The results show that the system can meet demand for precise feeding of aquatic products.

关 键 词：水产养殖 图像矩 单目视觉测距 嵌入式系统 

分 类 号：S951.2[农业科学—水产养殖]