Research On Intelligent Traceability Of Air Pollution Based On Multi-source Remote Sensing Information

Atmospheric pollution is a pressing issue that poses a significant challenge to the global environment,directly impacting human health and ecological stability.To effectively address air pollution,it is imperative to manage pollutant emissions,which not only aids in pollution prevention and control,but also enhances management efficiency.With the rapid advancement of remote sensing technology,remote sensing information has provided novel insights for air pollution management.This study utilizes multi-source remote sensing information to investigate the intelligent tracing of air pollutants,trends in pollutant distribution,and analyses of pollutant source emissions.We establish an intelligent traceability model of air pollutants and estimate the scale and emissions of pollutant sources.The primary research focus of this paper is outlined below:1.Comparative analysis of atmospheric pollutant inversion methods,atmospheric dispersion models,and swarm intelligence algorithms from the perspective of atmospheric pollutant traceability.The intelligent traceability model of atmospheric pollutants was developed using the Gaussian plume model combined with the firefly algorithm with an automatic step size.The firefly algorithm was selected based on factors such as algorithm complexity,global search effect,search effect on multi-peak function,and algorithm scalability,enable the model to combine pollutant concentration distribution information and meteorological information to perform intelligent retrieval of air pollutants.2.Through ground station monitoring data and GDAS data,the sources of pollutants in Hangzhou city were identified using backward trajectory clustering and potential source analysis,and the location and number of potential air pollution sources in Hangzhou were initially determined using satellite pollutant concentration images for particle processing and clustering based on the mean drift algorithm,and the results showed good agreement with the locations of officially announced air pollution enterprises.3.Two types of pollution sources of Gaussian plume model were simulated,followed by traceability verification through the dispersion simulation results of the atmospheric pollutant traceability model.The average image deviation for the positioning of the two types of pollution sources was 0.11 image elements.The improved dark pixel method was used to infer the aerosol concentration in satellite images,and a polluting enterprise with multiple surface sources was selected for traceability verification in a real environment.The final localization success rate was found to be 83.3% with an average deviation of 0.74 pixels,thereby validating the accuracy of the model.4.Analysis of the spatial and temporal characteristics of three pollutants in mainland China using satellite pollutant concentration data from 73 domestic cogeneration enterprises at the location of their emission outlets.The enterprise emission relationship was studied,and the pollutant emission inventory was constructed based on different power units by combining the official enterprise scale data and annual emission data.The study estimated the scale and annual emission of a cogeneration enterprise in Chengdu with a deviation of 10% from the total power of the actual enterprise units.