Inversion Of Aerosol Optical Depth In Beijing-Tianjin-Hebei Region Based On GF4 Satellite Data

Aerosol optical depth is characterization of aerosol concentration and the quantity of atmospheric turbidity,and currently realizing the inversion of atmospheric aerosol optical depth by means of remote sensing is one of the important means of atmospheric monitoring and management,but the method choice according to the characteristics of the satellite sensor imaging and band and conform to the region characteristic aerosol type has been the emphasis and difficulty in remote sensing inversion aerosol optical depth.Based on this,this paper inverts the aerosol optical depth in the Beijing-tianjin-hebei region by focusing on the characteristics of gaofen-4 satellite sensor and aerosol types.The main research contents of this paper are as follows:In view of the defect that the traditional dark pixel method cannot be directly applied to gaofen-4 data,the distribution and linear relationship of equivalent surface reflectance in red and blue wave bands in satellite data are obtained through experiments and researches in this paper.The processing processes of gaofen-4 image,such as orthorectification and radiometric calibration,are designed,and the features of dark pixels are extracted.The influence of input parameters in 6S model,such as geometric parameters,surface reflectance and aerosol model to the inversion effect of aerosol optical depth are studied.Using self-organizing feature mapping neural network and particle clustering algorithm,cluster analysis is carried out on the measured samples of local aerosol characteristics.The proportion and half-life of aerosol clustering results are analyzed,and three aerosol types are extracted and their corresponding look-up tables are established by 6S model.In order to study the results of AOD inversion of different aerosol types,the aerosol products of aerosol is conducted to consistency analyse with the aerosol products of MODIS,and validated by AERONET aerosol results respectively,whose applicability and characteristics are evaluated by comparing the correlation coefficient,relative error and root mean square error.Results show that the fine particles of C4 type aerosol satisfy the region characteristics of aerosols,summer,autumn season and has strong sensitivefeatures with the changes of the aerosol optical depth value,which also show that the efficiency of clustering methods and the clustering effect.