Application Of Multi-source Satellite Water Color Remote Sensing Based On Improved Color Matching Function Algorith

The color matching function makes it easy to synthesize color triple stimulus values using remote sensing multispectral data,but the current application of the color matching function is still limited by the small number of sensor band settings,large band intervals and uneven band settings,resulting in the color shift of the synthesized true color images and the inversion of the obtained water color index（Forel-Ule Index,FUI）cannot characterize the true water body.The inverse Forel-Ule Index（FUI）cannot characterize the real water body color.To solve the above problem,the color matching function algorithm is improved in this paper,and the color matching model of the target sensor is obtained by combining the true values of R,G,and B stimuli calculated by the color matching function of Hyperspectral Imager for the Coastal Ocean（HICO）,which has abundant channel settings in the visible range,with the waveband conversion method and Back Propagation（BP）neural network.The true color image synthesis models for Landsat 8 OLI,Terra MODIS,and Himawari-8 AHI and for Ocean Color Climate Change Initiative（OC-CCI）of ESA are obtained by using L1b and L3 remote sensing reflectance data as input data,respectively.Initiative（OC-CCI）climate state fusion data and the water color index FUI inversion model of Terra MODIS to demonstrate the comparability and application potential of this improved algorithm across multiple source sensors.The quality of the true-color images synthesized by this algorithm is evaluated by comparing it with the images synthesized by the currently used three-waveband method.The FUI inversion dataset is spatially and temporally matched with the real measurement dataset to verify the accuracy of the FUI inversion model in this paper.The correlation between chromatic parameters such as FUI and chromatic angle and water quality parameters were investigated,and the inversion model of chlorophyll concentration based on chromatic angle and chromatic distance and the evaluation model of eutrophication status of water bodies based on FUI were established.The conclusions are as follows:The correlation coefficients of the synthetic model of true color images of each sensor were above 0.99,and the model training results were good.Comparing and analyzing the true color images synthesized by the traditional three-band method and the model of this paper,we calculate the four objective evaluation parameters of mean value,standard deviation,average gradient and information entropy,and plot the changes of R,G and B in the cross-section of each feature type,combined with the subjective analysis of the true color images and color histogram,the results show that the improved algorithm of this paper effectively corrects the color bias of the three-band synthesized images,and the color and The results show that the improved algorithm can effectively correct the color bias of the three-band composite images,and the color and information of the features can be expressed more abundantly.Comparing the R,G,and B values of different feature types in the cross-section of the two algorithms,it is found that the same feature types with different sensors have the same trend of color tri-stimulus values,and the color of the images calculated by the model is more similar for different sensors.Based on the FUI long time series datasets of global oceanic waters and Chinese offshore,the spatio-temporal matching with the measured data is verified,and the algorithm of this paper has high accuracy.The spatial and temporal distribution and variation trends of global and Chinese offshore waters are analyzed.On the spatial scale,the FUI increases from low latitude to high latitude and decreases from near-shore to far-shore.On the interannual scale,except for the FUI of the oceanic water bodies from the equator to the Tropic of Cancer and the high value area along the East Siberian shelf,the FUI of all the offshore oceanic water bodies and high value areas showed an overall increasing trend;the FUI of the Bohai Sea and the estuary of the Yangtze River in the Chinese offshore region showed a decreasing trend,while the FUI of the sea area south of the Shandong Peninsula and the offshore area of the East China Sea showed an increasing trend.On the seasonal scale,the nearshore high value area is greatly influenced by seasonal changes,with the smallest FUI in summer and the largest FUI in winter,while the low latitude and offshore oceanic waters are little influenced by seasonal changes.Taking Chl-a as an example,the correlation between FUI and Chl-a was investigated,with the increase of FUI,Chl-a first increases and then decreases,and finally increases again gradually at the 15≤FUI place.The inverse model of Chl-a was established by adding the chromaticity distance d² and combining with the chromaticity angleα,and the upper limit of Chl-a obtained from the inversion could reach 24.18 mg·m^-3.Finally,the nutrient status index method based on chlorophyll concentration was chosen to establish the FUI-based water body eutrophication evaluation model,which realized the judgment of water body nutrient status by using water body color.This study explores the application of color matching improvement algorithm in true color and water color extraction,and demonstrates the stability and comparability of the algorithm on different sensors,which helps to realize the monitoring research on the ground and water bodies using multi-source sensors and meet the monitoring requirements of different levels of spatial and temporal resolution.