Water Quality Online Monitoring System Basedon Video Analysis

With the rapid development of society,the problem of environmental pollution has become increasingly severe,especially the water pollution.An effective water quality monitoring mechanism is an important part of water quality management.1.Two types of water quality monitoring methods are analyzed and compared:physical and chemical detection methods and biological detection methods.The principles and advantages of biological detection methods are introduced.The existing water quality monitoring systems are analyzed at home and abroad,and the principles and existing problems are stated.2.A water quality monitoring platform is designed.The hardware part simulates the simplest water quality monitoring environment.The zebrafish is used as the indicator organism.The copper sulfate solution is a toxic pollutant,simulates water pollution,and monitors fish movements in real time through the camera.The software part performs online analysis and processing on the video image sequence collected by the camera.A GUI interface for water quality monitoring is designed.The monitoring personnel can perform online observation of changes in the fish movement data simply by performing a simple operation on the GUI interface.Since the water quality monitoring system performs analysis and processing on the video image and the detection result is greatly affected by the light conditions of the monitoring environment,the video image of the input system is evaluated using the GWH-GLBH algorithm,and the light conditions of the monitoring environment are configured.3.Computer image processing technology are combined with water quality monitoring.In order to eliminate the influence of complex background on the extraction of moving target parameters,Seg algorithm is used to extract the foreground moving targets based on the saliency measurement algorithm and the conditional random field model,which can achieve better foreground extraction results.In order to analyze fish movement characteristics under different water quality conditions,fish movement speed,rotation angle,spatial standard deviation,and body color parameters were quantified and extracted,and fish movements are comprehensively evaluated from various aspects,so that analysis water quality status via the fish movement characteristics.4.The machine learning theory are combined with the water quality monitoring system and the classification model based on the LSTM neural network is used to classify the fish movement characteristics data under different water quality environments.The method used in this research system can achieve more accurate multi-levels result,and the quickly early warning speed.It has a high reference for the sudden situation of the water quality environment.