Research On Monitoring System And Key Technologies Of Rice Growth Environment Based On Internet Of Things

China is a large agricultural country with the highest rice production and consumption in the world.The Rice production has great significance to maintain stability of nation and society.At present,China’s rice production per mu yield and utilization rate of water,fertilizer and pesticide are still much higher than those of agricultural developed countries.It has been proved to be one of the effective ways to improve the production and resource utilization by collecting the information of rice growth environment which is used to direct agricultural production.The technology of agricultural Internet of things takes the application of information perception equipment,communication network and intelligent information processing technology as the core,through the scientific management of agriculture which can make rational use of agricultural resources,improve the ecological environment,reduce production costs and increase the production and quality of agricultural products.At present,the application of Internet of things technology in rice agriculture is faced with the problems of small network coverage,limited monitoring items,high transmission cost,short power supply time and low degree of intelligence,which limit the popularization and intelligent degree of rice growth environment monitoring.With the rapid development of low-power Wan,edge computing,artificial intelligence and other Internet of Things related technologies,which provide more perspectives to solve these problems.Therefore,combined with the Internet of Things and related technologies,this paper constructs a wide coverage,low-power and intelligent Internet of Things monitoring system which is suitable for monitoring the rice growth environment,and does some research in the key technologies of the system,which can be more comprehensive,accurate and real-time to master the environmental factors of rice growth,so as to better guide rice production.(1)The framework of Internet of Things for rice growth environment monitoring based on edge computing was studied.An Internet of Things framework for rice growth environment information collection with low power consumption,long transmission distance and integrated pest monitoring function was designed.The heterogeneous integration of high-definition image transmission and low-power Wan realized the functions of pest monitoring and rice growth environment information collection under the layout of low-power Wan.NS3 software is used to simulate the situation of laying a large number of sensor monitoring nodes in rice growing environment.The comparison of the advantage parameters combination range about network transmission quality under the two ACK mechanisms is obtained.This paper also gives the parameter combination range and suggestions to improve the network.(2)An on-line monitoring gateway of rice growth environment based on edge computing technology was designed.On the basis of supporting LoRaWAN and802.11 g heterogeneous networking,the architecture of functional and data communication of edge computing gateway are designed.Through virtualization container technology,LoRaWAN server,pest recognition and counting algorithm and data fusion algorithm are packaged into image.In the field of agricultural production,the edge computing mode of integrated operation with multiple functional modules is formed.The middleware of edge message is used to standardize and customize the data transmission between each functional module and between cloud and edge module.Through the actual test,the gateway can realize the function of each module at the same time.In 500/1000 concurrent stress tests per second,the average load is0.22/2.99 and the utilization rate of system resources is stable.Data transmission is also stable and reliable by using edge message middleware.The success rate of data transmission in field test is 99.1%.(3)A high accuracy online multi-sensor data fusion scheme is designed.Through the improved algorithm,the online data fusion test of the uploaded sensor data is carried out.Compared with the traditional fusion algorithm,the variance is reduced by about 25%,which effectively improves the accuracy of the obtained data of rice growth environment factors.(4)The artificial intelligence algorithm for on-line identification of rice pests was studied.Automatic image preprocessing is used for the collected images of rice pests.Traditional image processing techniques such as image enhancement and image segmentation are used to optimize the image quality.An online pest identification and counting method is proposed.Based on the deep learning algorithm of artificial intelligence,the on-line identification and counting of pest images uploaded by the pest monitoring node are completed under the TensorFlow framework,and after the field test the recognition accuracy reaches 89%.The digital output of image recognition results greatly reduces the pressure of Internet of things transmission and cloud computing.(5)The cloud platform of rice growth environment monitoring data management was constructed.The data collected by all the monitoring nodes are displayed dynamically and visually.According to the collected information and the expert system,the corresponding opinions are given to guide the rice production.