Aerosol Optical Depth Retrieval Over Urban Areas Using GF-4 PMS Data

GF-4 PMS satellites have the characteristics of large monitoring range,large width,fast response,and various observation and imaging modes throughout the day,which can provide powerful protection for aerosol optical thickness inversion.However,due to the high spatial resolution of GF4-PMS,short-wave infrared bands near 2.1 μm are still missing,and urban areas are high reflectivity areas,making it difficult to accurately determine the surface reflectance.The GF-4 PMS sensor has problems of insufficient detection information,difficulty in registration,and complicated underlying surfaces and high reflectivity in urban areas.Comparative study of aerosol inversion algorithms in urban are-as,the most use of the dark target algorithm combined with surface reflectance database to support the retrieval of aerosol optical thickness in urban areas.The basic idea is: using the percentage matching method of the surface reflectance database to divide the urban area into the dark target area and the bright target area,and use the dark target algorithm to achieve the retrieval of the aerosol optical thickness in the dark target area.Use MOD09 A1 in the bright target area.Data creation database supports inversion of aerosol optical thickness.The Beijing area was selected as the experimental area and the algorithm experiment was conducted.To verify the inversion accuracy,the results of the MOD04 aerosol produc t show that the verification results show that the aerosol optical depth in the urban area a nd the aerosol aerosol optical density of the MODIS land are inverted using this algorith m.The consistency is higher,the correlation of the overall verification is greater than 0.78,the correlation of the verification of the random sampling point is greater than 0.96,an d most of the inversion values are within the MODIS error line.The overall correlation b etween the inversion results obtained by the proposed algorithm and ground-based observat ions is relatively close.The correlations between the inversion results and the ground-base d observations are better than those of the ground observations.The correlation coefficient is better than 0.96.However,the overall GF-4 inversion results are higher than the grou nd observations.