鱼类应激行为作用下的水质视频监测分析系统

(水声通信与海洋信息技术教育部重点实验室(厦门大学),福建 厦门 361005)

生物监测; 目标识别; 目标跟踪; 计算机视觉

The Video Monitoring System of Water Quality Based on Stress Behavior Analysis of Fish
HUANG Yifan,CHEN Xin,YUAN Fei*

(Key Laboratory of Underwater Acoustic Communication and Marine Information Technology,Ministry of Education,Xiamen University,Xiamen 361005,China)

biological monitoring; target recognition; target tracking; computer vision

DOI: 10.6043/j.issn.0438-0479.201609010

备注

由于传统的理化分析法无法满足水质实时监测的要求,采用生物监测法,基于视觉化的计算机监测系统分析斑马鱼的应激行为,主要运用RGB颜色空间下的红蓝分量差分法实现运动目标检测,然后进行二值图像标记,并提出了一种最大重合面积跟踪算法.实验按污染时间顺序分为正常情况、污染初期及污染末期3种状态,以量化分析鱼群的各项行为特征参数; 基于统计学分析选取出可以监测水质变化的参数指标; 最后联合多维参数在两个过渡阶段实现分级报警.

Because the traditional physical and chemical analysis cannot meet the real-time requirement,the biological monitoring method is used in this paper in which we analyze the stress behavior of zebrafish based on the visual monitoring system of computers.In fact,we primarily use the difference of red and blue components in RGB color space to detect moving targets,then tag binary images obtained,and a tracking algorithm of maximum overlap area is proposed.In the experiment,there are three conditions according to the pollution time,and we quantify the behavior characteristic parameters of fish in normal,pollution-early,and pollution-later situations.And then we select some parameters which could monitor the water quality based on statistical analysis.Finally,a hierarchical alarm system has been accomplished in two transition stages by combining multidimensional parameters.