| With the rapid development of social economy,the risk of water pollution is increasing.As a ’ desert reservoir ’ in China,Haibowan Reservoir is a typical area for water quality monitoring.Suspended sediment concentration is an important parameter to measure water quality.The use of inland water color remote sensing technology can overcome the shortcomings of traditional sediment measurement methods,and provide a new technical way for the dynamic monitoring of suspended sediment concentration in the reservoir area.In this thesis,based on the measured hyperspectral and suspended sediment data,the optical characteristics of suspended sediment concentration in the Haibowan reservoir area are analyzed.Combined with synchronous multi-source remote sensing data,the remote sensing inversion estimation model of suspended sediment concentration and apparent optical quantity is established.Based on the water mapping algorithm,the long-term water area of Haibo Bay Reservoir was extracted,and the trend surface analysis method was used to analyze the change trend of water area.Finally,the optimal inversion model was applied to analyze the spatial and temporal distribution characteristics of suspended sediment concentration in Haibo Bay Reservoir in the typical period of water area change.Main research results and conclusions:1.The water body in Haibowan reservoir area conforms to the spectral characteristics of type II water body.The change trend of spectral reflectance curve in May and September is basically the same,and there are two peaks.The first main peak is about 570 nm,and the second main peak is about 820nm;In the range of 350-900 nm,the reflectivity of each band increases with the increase of suspended sediment concentration,and the increase position of different bands is basically consistent with the position of reflection peak.2.The band ratio and first-order differential model were constructed based on the measured spectral data;based on multi-source remote sensing data,a single band,band ratio and band combination model are constructed.Then,four evaluation indexes are used to verify and analyze the accuracy of the suspended sediment inversion model.The study shows that the fitting degree of band ratio model of Landsat image data is the best,and its determination coefficients in May and September reach 0.8678 and 0.8712.3.The pixel error rate of the water area extraction accuracy of Haibowan Reservoir is4.36%.The interannual variation trend of water area in the reservoir area is studied.In the spatial and temporal dimension,the overall water area decreased first and then increased from 1990 to 2021,with a total increase of 79.64km~2,and there are different contraction and expansion trends in different periods.Due to the construction of water conservancy projects in 2013,the upstream inflow was blocked,resulting in the development period of water area from 2013 to 2021.4.According to the variation characteristics of water area in the long time series reservoir area,the optimal inversion model of suspended sediment concentration was applied to Landsat remote sensing images in May and September 2013-2021 to analyze the spatial and temporal distribution characteristics of suspended sediment concentration in the reservoir area.From the perspective of time variation,the suspended sediment concentration in the reservoir area showed a trend of first decreasing and then increasing from 2013 to2021,and the variation range of concentration was between 10mg/L and 300mg/L.From the perspective of seasonal distribution,there are obvious dry and rainy seasons,and the suspended sediment concentration after flood is higher than that before flood.From the spatial position,the suspended sediment concentration in the right bank is higher than that in the left bank,and the concentration value is the highest in the front of the dam,and the concentration value decreases along the flow direction. |
PDF Full Text Request