The Design Of The Grain Monitoring And Safety Evaluation System Based On Multi-view Learning

As an important strategic material for national survival and development,grain is closely related to economic development,social stability and people’s happiness.With the development of science and technology in China,the domestic grain storage situation has been continuously improved,and the development of grain condition monitoring system has gradually moved towards modernization and intelligence.However,most domestic granaries still monitor and manage the granaries in the traditional manual way.Compared with the modern intelligent grain monitoring system,the traditional grain monitoring system has the following problems: 1.The monitoring of grain situation data is single,the grain safety status of granary cannot be accurately identified,and the technology of grain situation analysis is backward;2.The communication cable wiring is complex,and the communication distance is short and the power consumption is high;3.The efficiency of granary monitoring and management is not high,and the intelligent control of grain conditions is low,and most of them are based on manual decision-making control;4.The system functions are imperfect,lacking real-time grain data recording and storage functions,and needs to be optimized in data display and interface operations.In view of the above problems,this paper designs a grain condition monitoring and safety evaluation system based on multi-view learning.The system adopts the mode of terminal collecting data,concentrator processing and transmitting data,cloud server storing data and man-machine interface monitoring.The whole system collects the data that affects the grain safety status of the granary by using a variety of smart sensors in the terminal equipment,and transmits the collected grain data to the concentrator through wireless Lo Ra communication to process the data,and the PSO-LSSVM model is used on the concentrator to conduct multi-view training and learning on the collected grain data and evaluate the food security status.On the one hand,it issues commands to the terminal in a timely manner to control the factors that affect grain security when evaluating and judging the food security situation;on the other hand,the concentrator uses wireless Wi-Fi to connect to the Internet and uploads the grain condition data to the One NET cloud server to store and record the data and display the man-machine interface.It is in this mode that the system realizes long-distance monitoring,which greatly improves the flexibility and scalability of the grain monitoring system,and effectively solves the shortcomings of the current grain monitoring system,such as single monitoring data,insufficient grain analysis ability and low level of intelligence.Combined with the PSO-LSSVM optimization algorithm,this system performs multi-perspective data analysis and processing on the detected granary environmental parameters（granary air temperature,granary air humidity,grain temperature,water content of grain and water,granary CO₂ concentration）.And it also evaluates the grain safety status of the granary and grades the safety status,and outputs the granary grain safety assessment status level（I-safe,II-warning,III-dangerous,IV-unsafe）.The simulation results show that the PSO-LSSVM algorithm model has a classification accuracy rate of98.75% for the granary grain safety assessment state,and the classification performance is better.Through data testing and experimental inspection,the system can ensure the stable operation of the multi-view learning grain monitoring system.The terminal and concentrator have high data processing efficiency and basically meet the design requirements of the expected grain monitoring system.While improving the management efficiency of the granary,the safety of the granary is greatly improved,and new technical means and scientific basis are provided for the practical application of the granary grain condition monitoring and management system.