Studies On Water Quality Monitoring System For Aquaculture Based On Information Fusion

With the rapid development and the adjustment of the structure of China’s aquaculture. aquaculture water quality monitoring plays a decisive role in increasing production, reducing costs, improving economic efficiency and preventing diseases in fish. For complex and non-linear characteristics of farmed aquatic environment, this study is based on an information fusion method which is applied to aquaculture water quality monitoring, in order to realize accurate monitoring of the parameters of water quality and use of information fusion model for water quality prediction and evaluation purposes.(1) the thesis describes the status of water quality monitoring of key technology research and development at home and abroad, on the basis of that proposes aquaculture water quality technical indicators and a distributed computer control system based on Wireless sensor, and analysis on highlights and features of each subsystem technology(2) designed information collecting on the child nodes of aquaculture water quality based on wireless sensor, achieve multi-sensor parameter acquisition of temperature, dissolved oxygen, pH and turbidity. Meanwhile, in order to ensure the accuracy,some research has been done including the nonlinear studies of temperature and turbidity,the compensation of dissolved oxygen and pH to temperature.(3) the paper focuses on the application of information fusion method in aquaculture water quality monitoring. Firstly, the paper designed a two-stage information fusion model:The first level is the similar sensor fusion problems, based on an adaptive weighted fusion algorithm to improve the detection accuracy of similar water quality parameters; And the second level is the heterogeneous sensor fusion problem, through building a fusion model based on neural network to achieve water quality prediction and quality evaluation functions. Secondly, this thesis presents an improved LMBP algorithm through analysis and comparison of the standard BP algorithm and traditional LMBP algorithm, The result shows that the improved algorithm has the fastest convergence under the premise of meeting the requirements of error.Lastly, test the model by test the data from breeding base, The result shows that model is reliable, the precision for prediction and assessment of water quality is high.(4)using C#programming language developed aquaculture water quality monitoring system client applications which is based on Visual Studio2010development platform and winform form. Software features meets the needs of the monitoring system and has some practical value.