| Poyang Lake is the largest freshwater lake in C hina,is also an important water source of the middle and lower reaches of the Yangtze River and the largest natural flood diversion and storage areas in the Yangtze River basin.Therefore effective monitoring of water quality and water levels in Poyang Lake is particularly important.Remote sensing has a great advantage in the monitor due to its large range,high efficiency,high spatial resolution and less access to information subject to conditions,but atmospheric correction is always the key factor limiting its opera tional application.Currently,atmospheric correction algorithm for class one water bodies has been more mature and atmospheric correction for inland water issues are not fully resolved.Based on this,taking Poyang Lake waters as the study area,2013 to 2015 Landsat OLI image as the data source,based on the measured data,explore the atmospheric correction of Landsat OLI images.In this paper,atmospheric correction of Poyang Lake water body is carried out by using case ? water body atmospheric correction algorithm based on clean water body,case ? water shortwave infrared algorithm,case ? water body atmospheric correction algorithm based on MODIS auxiliary data and iterative algorithm.(1)Determine the water level of each site by measuring the sediment concentration,obtain the spectral curves of the different water bodies on the uncorrected images,and use the statistical analysis to find the clean water body.In 2013,the clean water bodies are mainly distributed in the Poyang Lake area northeast corner of the military and civilian reservoir and the eastern part of Poyang Lake.Using the known clean water,calculate the atmospheric calibration parameters,and then complete the atmospheric correction.(2)Based on the assumption that the Landsat OLI image SWIR1(effective wavelength is 1609nm)and SWIR2(effective wavelength is 2201nm)from the hypothesis that the water radiation is zero,the aerosol scattering of these two bands is obtained,and get the atmospheric correction parameters,complete atmospheric correction.(3)Based on the assumption that the fifth wavelength band(effective wavelength is 1240nm)and the sixth band(effective wavelength is 1640nm)of MODIS image is zero,the atmospheric correction parameters are obtained and the atmospheric calibration parameters are applied to the Landsat OLI images which atmosphere is consistent with the MODIS image,and then complete the Landsat O LI image of the atmospheric correction.(4)Using the relationship between the reflectivity of water body and the inherent optical properties of water,combined with the measured chlorophyll concentration,an iterative algorithm based on shortwave infrared algorithm optimization is realized.When the two iterations in the shortwave infrared band(effective wavelength is 1609nm)are less than 5%,the iteration is stopped.The measured data from 2013 to 2015 and the results of four kinds of atmospheric correction were compared under five water bodies.The results show that the correctness of the correction based on the clean water algorithm is the best in the turbid zone,the medium turbid zone,the mild turbidity area,and the correction error of each site is not more than 20%.In the clean water area and especially the clear area,the correction effect based on the clean water body algorithm is the worst,in the other three methods,the iterative algorithm is the best.The relative error of the two sites in the special clear area in 2014 is 19.6% 18.28%;in 2015,two sites in the clean water area,the relative error of 14.5%,17.96%.According to the above conclusions,an iterative classification of image is realized,and the image is partitioned on the basis of this classification.Compared with FLAASH correction,the spectral reflection is closer to the measured data,indicating that the partition correction is feasible and Which provides a new idea for the study of atmospheric correction of Poyang Lake. |
PDF Full Text Request