Key Technology And Application Of Smart Fish Pond Based On Deep Learning And Fuzzy Logic

Water quality monitoring is an important link to ensure the healthy growth of fish.The traditional monitoring methods are time-consuming,laborious and easily disturbed by environment,this paper introduces the theory of depth learning and fuzzy control into the field of water quality monitoring.Deep Learning is a machine learning method based on neural network.It can extract and analyze the key features of data by learning the complex patterns of data.Fuzzy control is a kind of control method aiming at fuzzy and uncertain information.It realizes the control and optimization of complex system by fuzzy set,fuzzy logic and fuzzy reasoning.The main contributions of this paper are as follows:Firstly,propose a deep learning-based intelligent fish pond water quality monitoring technology.Use deep learning networks based on Long Short-Term Memory(LSTM),Gated Recurrent Unit(GRU),and Attention Mechanism to predict water quality.Adopt a multi-layer LSTM network structure,use GRU to handle the time series factors of the data,and predict water quality parameters by sequentially representing the input LSTM-processed feature vectors in GRU.By comparing the model and water quality prediction effects,it is proved that the efficiency of water quality prediction mixed with LSTM,GRU,and Attention and the stability of the control framework are improved.Secondly,design the control algorithm of fish pond monitoring node based on fuzzy logic.In this paper,a fuzzy logic model is established by collecting data from sensors,and fuzzy reasoning is carried out.According to the control signal obtained by de-fuzzification,the outputs of control,ph regulator and oxygen supply are obtained,the final realization of PH and do automatic control.Through fuzzification,anti-fuzzification and experiments,the reliability of the fuzzy algorithm is verified.Finally,develop the intelligent fishpond management platform based on the three-layer structure of Internet of things.In hardware,the monitor node and sink node are designed,and the communication between sensing layer and transmission layer is realized by Zig Bee wireless communication technology and 5G mobile communication technology.On the software side,the application layer development is realized by using One NET third-party cloud platform and Android mobile terminal.The feasibility and reliability of the system are verified by test experiment.The test shows that the platform has perfect function and stable performance.In this study,the structure of deep learning network is used to realize the automatic learning and feature extraction of underwater data,to achieve the efficient processing and prediction of water quality data of fishpond,and to adjust parameters adaptively,to improve the performance and robustness of the system,to process the uncertain information in the water quality data of fishpond by fuzzy control,to avoid misjudgment or miscontrol of the system due to incomplete or inaccurate data,improve the reliability and stability of the system.This study can provide some theoretical support for the construction of aquaculture information platform.