Study On Typical Saline Region By Remote Sensing Monitoring In The Yellow River Delta

Soil salinization in the Yellow River delta is serious,posing a threat to the stability of the ecological environment,becoming the main factor restricting the development of the regional economy.The development of satellite re mote sensing technology provides a fast and convenient mean for monitoring large area’s soil salinity.Soil salinization monitoring mainly based on spectral features for the quantitative extraction of soil salinity information.Soil spectral characteristics are influenced by atmospheric correction,soil moisture,vegetatio n coverage.So it needs further stud y to reduce their impacts on the accuracy of extracting salinization information.Hyperspectral data sets up a new way for the study of quantitative relationship between soil salinity and spectral characteristics.Application of remote sensing technology,investigating and mapping saline condition based on measured soil spectrum and Landndsat-8 image data,poviding the possibility for fast,accurate and comprehensive monitoring of salinization.Based on the typical saline region in the Yellow River Delta area,Kenli County,Dongying city as the research object,getting the measured oil hyperspectral data providing by AvaField field spectrometer,the data of soil salt content,the remote sensing image data.Using the partial least squares method with the analysis of the correlation between the bands of measured soil hyperspectral reflectance data and the soil salt content.Using full wave bands and the extracted sensitive wave bands to modeling and analysing respectively,finally establishing the estimation model of soil salinity.Using the multi spectral data of Landsat-8 images to extract soil salt content,first introducing the theory,method and process of Landsat-8 radiometric calibration based on Flaash,then extracting the sensitive bands by spectral diagnosis index,with which establishing the multiple linear regression model for the inversion of soil salinity based on spectral transform.Discussing the design of BP neural network model on soil salinity remote sensing inversion,programming and testing the model accuracy in the MATLAB environment.BP neural network is proved that it has feasibility and advantages in the simulation and prediction of soil salt content,finally made the spatial distribution map based on soil salt content inversion model.Revealing that Landsat-8 satellite multi-spectral data has the potent to detect soil salinization.Through digging the spectral information of sa line soil,a new academic attempt to quickly obtain saline soil salinity information,but the accuracy has yet to be further improved.By means of remote sensing technology and GIS analysis,studying the salinization degree,type and distribution in the saline area,in order to achieve the purpose of monitoring of soil salinization rapidly and accurately.