基于生物水质监测的鱼体运动状态检测系统研究

发布时间：2018-01-02 01:01

本文关键词：基于生物水质监测的鱼体运动状态检测系统研究　出处：《西安邮电大学》2017年硕士论文　论文类型：学位论文

【摘要】：在水污染防治中,智能高效的水质监测系统在水环境实时监测及有效治理中发挥着至关重要的作用。生物水质监测克服了理化监测只针对特定指标的局限性及连续取样的繁琐性,能够直观、快速、综合地反映水质指标。鱼类作为重要的水质指示生物,对于水环境参数的变化较为敏感,其运动特征直接表征水质污染状况。监测鱼类运动状态参数的实时变化规律,将为进一步研究鱼类水质指示特性及其所处水环境水质变化状况奠定基础,对于提高水质监测的智能化水平具有重要意义。本文基于2014年陕西省教育厅产业化培育项目(水质远程分析科学决策智能化环保系统研制,项目编号:14JF022)完成了生物水质监测子课题,提出了一种基于生物水质监测的鱼体运动状态检测方法,建立水质监测模型,并开发了一套适用于生物水质监测的鱼体运动状态检测系统。本文所做的主要工作如下:1.提出了一种鱼体运动目标视频图像检测方法。采用统计学背景建模算法快速建立背景模型,通过背景差分算法将鱼体目标与背景分离,得到灰度化鱼体目标图像;采用最大类间方差算法(Otsu)获取自适应分割阈值,通过图像分割得到二值化鱼体目标图像;提出形态学滤波与面积比较法相结合的滤波算法有效滤除噪声并完整得保留鱼体目标;通过连通区域质心检测方法获取目标质心。2.根据鱼体目标区域质心和面积特征,建立了基于特征的目标追踪模型,实时追踪鱼体目标;获取鱼体目标尾点、头点等特征点坐标,建立鱼体运动速度、加速度、摆尾频率、转弯频率及撞壁频率求解模型;根据鱼体运动状态各参数水质语义特征,建立水质监测模型。3.根据以上算法及模型,基于LabVIEW设计了生物水质监测系统软件,完成了系统简介、视频采集、目标检测、记录查询及水质评价模块设计,实现对摄像头采集及加载已存储的鱼体运动视频实时处理,并进行水质超标报警。4.以红鲫鱼为实验鱼种,以改变水质pH为例,设计实验方案,测试系统性能。分别采集过酸(pH=3.5,4.5,5.5)、过碱(pH=9.5,10.5,11.5)以及适合红鲫鱼生长的最佳范围(pH=6.5,7,7.5,8)每种水质下4800帧(20min)鱼体运动视频,通过系统分析处理,实现水质过酸、过碱超标报警。以pH=10.5为例,第2986帧(12分27秒)时系统水质超标报警。5.对系统误差进行分析,主要误差在于目标检测过程中目标的丢失与增加,分析了各pH值对应的水质中鱼体目标检测的准确率,其平均准确率在0.91以上。上述工作的完成,实现了通过检测鱼体运动状态参数实时监测水质变化信息,达到了项目的立项要求,同时对于进一步研究鱼类生物水质监测具有重要的借鉴意义。
[Abstract]:In the prevention of water pollution. Intelligent and efficient water quality monitoring system plays an important role in real-time monitoring and effective treatment of water environment. Biological water quality monitoring overcomes the limitation of physical and chemical monitoring only for specific indicators and the complexity of continuous sampling. Fish, as an important indicator of water quality, is sensitive to the change of water environmental parameters. Monitoring the real-time variation rule of fish movement state parameters will lay a foundation for further study of fish water quality indication characteristics and water environment water quality changes. It is of great significance to improve the intelligent level of water quality monitoring. This paper is based on the industrialization cultivation project of Shaanxi Provincial Education Department in 2014 (the development of intelligent environmental protection system for water quality remote analysis scientific decision making). Item #: 14JF022) completed the sub-task of biological water quality monitoring, and put forward a method to detect the moving state of fish body based on biological water quality monitoring, and set up a water quality monitoring model. A set of fish motion state detection system is developed for biological water quality monitoring. The main work of this paper is as follows:. 1. A video image detection method for moving object of fish is proposed, and the background model is established quickly by using statistical background modeling algorithm. The background difference algorithm is used to separate the fish object from the background, and the grayscale image of the fish object is obtained. The maximum inter-class variance (Otsu) algorithm is used to obtain the adaptive segmentation threshold, and the binary fish object image is obtained by image segmentation. A new filtering algorithm, which combines morphological filtering with area comparison, is proposed to effectively filter noise and keep fish body targets intact. The centroid of the target is obtained by detecting the centroid of the connected region. According to the centroid and area characteristics of the target area of the fish, a feature-based target tracking model is established to track the target in real time. The coordinates of the characteristic points such as the tail point and the head point of the fish body are obtained, and the solving model of the moving speed, acceleration, pendulum tail frequency, turning frequency and wall collision frequency of the fish body is established. According to the semantic characteristics of the water quality of the moving state of the fish, the water quality monitoring model. 3. According to the above algorithm and model, the software of the biological water quality monitoring system is designed based on LabVIEW, and the system brief introduction is completed. Video capture, target detection, record query and water quality evaluation module are designed to realize the real-time processing of the video of the fish body which has been stored and captured by the camera. Taking the red crucian carp as the experimental fish species and changing the pH value of the water as an example, the experimental scheme was designed and the performance of the system was tested. And the optimum range for the growth of Carassius auratus is pH 6.5 ~ 7.5.The pH value of the red crucian carp is 7.5%, and the pH value of the red crucian carp is 7.5%. 8) 4800 frames / 20 mins of fish body motion video under each water quality, through systematic analysis and processing, the water quality is over-acid, and the excessive alkali alarm is realized. Take pH=10.5 as an example. In frame 2986 (12 minutes 27 seconds), the system water quality exceeds the standard alarm .5. the main error lies in the loss and increase of the target in the process of target detection. The accuracy rate of fish body target detection in water corresponding to pH value was analyzed. The average accuracy rate of fish body detection was above 0.91. The above work was completed. The real-time monitoring of water quality change information by detecting fish moving state parameters has been realized, which meets the requirements of the project, and has important reference significance for further research on fish biological water quality monitoring.
【学位授予单位】：西安邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：X835;S917.4;TP391.41

【相似文献】