Research On Regional Atmospheric Environment Monitoring System Based On Multi-sensor Fusion

Due to the imbalanced development of various regions,residential clusters and industrial parks are both producers and victims of atmospheric pollutants.With the rapid advancement of urbanization and industrialization,automobile exhaust,catering fumes,and coal-fired heating and corporate emissions will cause pollution to the atmospheric environment,which will have a huge impact on humans and the climate.At present,the assessment indexes of regional atmospheric environment mainly include six parameters concentration of air pollutants and air quality index,acquired and calculated regularly by the state control monitoring station.However,due to the characteristics of sparse deployment and timing sampling analysis of national control monitoring stations,it is inevitable to use monitoring data to assess regional atmospheric environment,and there are data redundancy and low timeliness,which is not conducive to the fine supervision of regional small-scale atmospheric environment.This thesis studies the regional atmospheric environment monitoring system under the above background.The main research contents are as follows:(1)Analyze the current status of domestic and foreign research on the atmospheric environmental monitoring systems and multi-sensors fusion.According to the application scenario of this system,design an atmospheric environmental monitoring system for small-scale areas based on the three-tier architecture of the Internet of Things.Use Long-distance LoRa technology establishes the monitoring network.The monitoring nodes in the perception layer and the LoRa gateway are designed in the software and hardware.At the nodes,the Free RTOS real-time operating system is used for multi-task scheduling to achieve the collection,fusion,and transmission.The gateway uses the embedded Linux system to realize the aggregation and transmission of node data.(2)By analyzing the data characteristics of the atmospheric environment monitoring nodes,and consider the situation that the atmospheric pollutant concentration changes slowly in small-scale areas under normal circumstances and data transmission is caused by timing transmission,comes up with an temporal correlation fusion algorithm(AMF-LR).It consists of adaptive median filtering(AMF)and linear regression prediction(LR),to achieve rapid response to significant changes in the environment and reduce data redundancy.(3)In order to realize the response of the emergency events in the monitoring area and the assessment of the overall air quality,based on the characteristics of similar changes in multiple nodes in the monitoring area and the high spatial correlation of neighboring nodes,an adaptive batching based on spatial correlation was proposed.The estimation fusion algorithm introduces the concept of local density in the density peak clustering(DPC)and gray correlation analysis(GRA)to detect abnormal data and alarms.The adaptive batch estimation fusion algorithm is used to obtain regional fusion values to improve regional air quality assessment capabilities.(4)To verify the performance of the monitoring system,tests node accuracy,system function,LoRa transmission effect,and fusion algorithm.Finally,summarizes the work done in this paper,and prospects the application and function update of the system.