Research And Application Of Crop Water Requirement Forecasting Model Based On Grey Neural Network

Estimate Crop water requirement is one of the critical aspects of irrigation management. To carry out scientific and efficient irrigation management, must accurately calculate and predict the Crop water requirement. With the rapid development of agricultural information technology for the construction of agricultural information provides a powerful technical support, system automation, intelligence as Crop water requirement predict future trends.Aiming at this problem have been studied on the basis of the theory of grey system theory and artificial neural network, the establishment algorithm prediction model, the use of grey neural network algorithm to predict the Crop water requirement, the model multi-dimensional data input meteorological factors, the use of Matlab tools to achieve crop water demand projections, and applied in Danzhou SongTao Irrigation Management Information System in Hainan Province.Firstly, the full text of the grey system theory and artificial through the network, through the relevant factors associated with meteorological and Crop water demand analysis to establish the grey neural network topology, the successful implementation of grey neural network model capable of predicting crop water requirement.Secondly, the use of C#and Matlab programming, using Matlab interface technology, the grey neural network prediction model compiled as COM components, to achieve a call the next.NET platform Matlab network prediction model, complete Crop water demand prediction algorithm integrated design.Finally, following the Crop water demand forecasting module intelligent interface development based on the principles, to carry out the prediction module functional design, the successful conclusion of the specific steps of the module functions and interface design, to provide guidance for crop irrigation.Simulation and test results show that the prediction of crop water demand curve and the measured curve fitting is high, based on grey neural network model predictions and actual crop water requirements mean absolute relative error of5.28%, high prediction accuracy, the realization of the crop accurate prediction of the basic functions of water for water-saving irrigation provides a new and effective method.