Development Of Agricultural Irrigation Area Information Monitoring System Based On Machine Learnin

With the development of society and the increase of population,the supply-demand relationship of water resources is becoming increasingly tense.60% of China’s water resources are used for agricultural production,and irrigation water accounts for 80% of agricultural water use.Agricultural irrigation areas are an essential component of agricultural production.Information monitoring and analysis of agricultural irrigation areas can provide a reliable basis for agricultural precision irrigation.This article was based on the Internet of Things technology to develop a low-power agricultural information monitoring device and data platform to obtain long-term and stable meteorological data from monitoring nodes.Using NB Io T technology as the communication solution and STM32 chip as the master control,collect information such as air temperature and humidity,soil temperature and humidity,wind speed,light intensity,and GPS in the irrigation area,and store the collected data on the Internet of Things cloud platform.Aiming at estimating evapotranspiration,which is crucial in the decision-making of irrigation scheduling,this paper proposes an estimation model of evapotranspiration based on regional division.The regional model of dividing meteorological stations with machine learning features estimates the reference crop evapotranspiration.Compared with the local estimation model of a single meteorological station and the generalized estimation model combining all station data,the regional model proposed in this paper has higher accuracy.For the problem of irrigation water level prediction,a multivariate water level prediction model incorporating meteorological data was also proposed in this paper.The water level data of irrigation reservoirs are analyzed in time series,the modal components of water level data are decomposed by the VMD decomposition method,and a VMDCNN-GRU multivariate water level prediction model incorporating evaporative meteorological data is established.Compared with the model with single water level variable,the performance is better and more stable in the multi-step water level prediction.Based on the above research,the software platform of the agricultural irrigation area monitoring system was built,and corresponding Web page have been deployed,and the system integration of irrigation area information monitoring,evapotranspiration estimation,and water level prediction was realized.The experimental results indicate that the system can monitor and analyze the information of agricultural irrigation areas,providing a reliable basis for agricultural precision irrigation.