| Aerosol is the main component of air pollutants and an important source of urban photochemical smog.Aerosols influence the earth-atmosphere radiation balance through direct and indirect effects.Spaceborne lidar has the advantages of high temporal and spatial resolution,day and night use,and wide detection area.It has become an important technical means to effectively obtain the temporal and spatial distribution of aerosol and cloud optical characteristics in a wide area and globally.At present,the aerosol vertical profile information of a large area is obtained using the data products provided by the US NASA CALIOP.In order to obtain more comprehensive aerosol vertical observation data and near-ground haze information,it is necessary to develop a spaceborne lidar.It monitors the environmental conditions of pollution within the region of my country,obtains the accurate temporal and spatial distribution of pollution information in China,and identifies and quantifies the concentration of pollutants in the region.The research work of this paper mainly includes two aspects.On the one hand,according to the requirements of spaceborne lidar detection,develop a scaled prototype of spaceborne lidar,and then carry out simulation research on the performance of spaceborne lidar;on the other hand,it is aimed at the measurement of spaceborne lidar.Signal characteristics,and carry out research on the quantitative inversion algorithm of aerosol optical characteristics and mass concentration in the boundary layer.In the field of simulation research on the performance of spaceborne aerosol lidar.According to the requirements of spaceborne lidar detection,a scaled-down prototype system was developed.The system adopts the same photon counting lidar detection system as the spaceborne lidar.Through the optimized design of the optical machine and structural modal analysis,it ensures high-precision alignment and maintenance between the outgoing laser and the receiving optics,and effectively solves the background light interference during the day.Realize the day and night observation ability of atmospheric aerosol in the height range of 10km.At the same time,based on the actual measurement data of the scaled-down prototype,by building a space-borne lidar to measure the effective signal,background signal and signal-to-noise ratio simulation,the focus is on evaluating the space-borne lidar under three typical weather conditions of clean,haze and cloud.Detection capability and vertical profile data quality,and from this,recommendations for the follow-up development of spaceborne lidar are given.In the research of spaceborne aerosol lidar data retrieval algorithm.First,a comparative analysis of the three meter scattering lidar inversion algorithms of Collis slope method,Klett method and Fernald method is carried out,and the uncertainty analysis of the Fernald inversion algorithm is emphasized.Secondly,according to the characteristics of the lidar signal,an adaptive threshold function denoising algorithm is adopted.By adjusting the threshold function,the useful signal can be retained to the maximum while denoising,and the signal-to-noise ratio of the measured signal of the lidar is significantly improved.At the same time,a segmented inversion of the aerosol optimization algorithm under the cloud is proposed.The differential zero-cross method is used to identify the height of the cloud top and the height of the cloud base.The cloud lidar ratio is reasonably selected through iterative inversion to achieve the vertical profile of the aerosol optical characteristics under the cloud Accurate inversion.Finally,by establishing a regression prediction model based on the acquired optical characteristics,meteorological elements and the measured particle concentration,the identification of the particle concentration is realized.The stepwise discriminant method is used to screen the characteristic variables,and linear regression,BP neural network and GA-BP neural network are used to construct an identification regression model to achieve the scoring prediction of particulate matter concentration.Six feature variables are selected to form the feature set through the stepwise discriminant method.At this time,R2 is the highest and the RMSE value is the smallest,which are 0.98 and 0.19,respectively.Through the regression model,it is found that the prediction error range of GA-BP is smaller than that of BP,and the prediction effect is better than that of BP method.The correlation index of the training set of the regression model is R2=0.904,the correlation index of the test set is R2=0.899,and the average prediction error is 7.122?g/m3,indicating that lidar can effectively monitor the distribution of particulate matter.It shows that lidar can effectively monitor the distribution of particulate matter.It proves that Lidar can be used as an effective and flexible instrument to collect particulate matter concentration data,especially for monitoring the spatial distribution of PM2.5 concentration in the atmosphere.In summary,the satellite-borne aerosol lidar scaled-down prototype can meet the observation needs.The satellite-borne analog signal cloud layer and pollution layer of this lidar have an obvious layer structure.The data products obtained by spaceborne aerosol lidar after orbit can be combined with meteorological element data to characterize the concentration of particulate matter,and realize the conversion of qualitative remote sensing and quantitative remote sensing. |
PDF Full Text Request