Research On Integrated Estimatation Of Water Color Parameters Based On Water-atmosphere Radiation Transfer Simulation

For the retrieval of water color parameters, the remote sensing images are usually corrected by atmosphere, and then the remote sensing information model is used to estimate the color parameters. In the atmospheric correction of inland case ? water, atmospheric corrections such as 6S model and MODTRAN model are often used. Atmospheric correction is often caused by the lack of satellite aerosol optical thickness, water vapor content and so on, which often affect the accuracy of atmospheric correction. Based on the water-air radiation transmission model, the aerosol thickness,chlorophyll, suspended matter and CDOM concentration multidimensional element look-up table were established,and the mechanism of water-atmospheric interface-atmosphere radiation was studied. The aerosol optical density, the concentration of chlorophyll a, the concentration of inorganic suspended matter and the absorption coefficient of CDOM440nm were simultaneously retrieved by aerosol - water element integrated remote sensing inversion algorithm. The algorithm proposed in this paper does not need to observe the atmospheric parameters synchronously, which improves the applicability of the model and the efficiency of the water color parameter inversion. The main contents and conclusions of the study are as follows:(1) Atmosphere - Water multidimensional element look-up table construction.According to the measured data of inland water bodies such as Taihu Lake,Chaohu Lake, Dianchi Lake, Dongting Lake, Hongze Lake and Three Gorges Reservoir Area, the boundary value of water - air radiation transmission simulation is established. The atmospheric information lookup table and water color factor concentration table are constructed respectively. In the atmosphere lookup table,the atmospheric top-level remote sensing reflectance look-up table based on the observed geometrical and aerosol optical thickness simulation at 550 nm was constructed for the Sentinel-2A satellite sensor using the 6S radiation transmission model. Among them, for the Sentinel -2A sensor, A band, there are 7 × 7 × 11 × 11 × 20 for the 118580 group of data; in the water color factor concentration lookup table, respectively, the establishment of the color element concentration of coarse lookup table and the color element concentration accurate lookup table, a rough lookup table, chlorophyll A concentration range is set to 0-150mg / m3, step size is set to 5mg / m3; inorganic suspended' particulate concentration range is set to 0-150mg / L, step size is set to 5mg / L; CDOM absorption coefficient range is set to 0.1-1.3 m-1, the step size is set to 0.2m-1, a total of 31×31×7=6727 data. In the exact search table, the concentration range of chlorophyll a is set to 20-100mg / m3, the step size is set to 0.1 mg / m3; the concentration range of inorganic suspended particulates is set to 20-100mg / L,the step size is set to 0.1mg/L; The coefficient range is set to 0.1-1.3m-1, the step size is set to O.1m-1,a total of 810 × 810 × 13 = 8529300 data.(2) Aerosol-water color parameter remote sensing integrated retrieval algorithm. First, to find the stable pixels in image, using the atmospheric information lookup table gets closest to the atmospheric path radiation and atmospheric transmittance, then combined color elements concentration lookup table, using two step searching idea, realize the integration inversion of aerosol optical thickness,chlorophyll,inorganic solids and CDOM440nm absorption coefficient. The first step, two step search, feature bands based on the determined by the spectral angle and spectral matching optimization method, first, color elements concentration rough look-up table, roughly determine groups of water concentration range; the second step, according to the scope of the first step of calculating the concentration of water group,and the spectral angle matching and optimization method is applied to color spectrum elements concentration accurately estimate color lookup table,element concentration, improve the calculation efficiency.(3) Algorithm accuracy test. The accuracy of the inversion was evaluated using the Sentinel -2A image acquired on July 23 , 2016. Firstly, the integrated inversion algorithm was used to calculate the chlorophyll a concentration,the concentration of inorganic suspended matter and the absorption coefficient of CDOM440nm in Taihu Lake. Secondly, the results of inversion were compared with the 10 sampling data measured by ground synchronization. The results showed that chlorophyll a The average relative error is 10.29%, the mean square error is 1.05, and the coefficient of determination coefficient of CDOM440nm is 0.367, the average relative error is 0.957, the average relative error is 10.26%, the mean square error is 1.05, the coefficient of determination is 0.367 , The average relative error is 41.64% and the root mean square error is 0.13.550nm. The aerosol optical thickness is measured by the measured value of AERONET in Taihu Lake Station. The measured aerosol optical thickness is 0.25, the matching result is 0.2, the measured value is one day aerosol The average thickness of the optical thickness, the matching results are ideal.