Remote Sensing Classification Mapping Of Saline-alkali Wasteland Grassland In Western Jilin Province

Soil salinization is caused by natural and human factors,which can inhibit the normal growth of crops and pose a serious threat to human existence.The degree of land salinization is directly related to the sustainable utilization of soil in arid areas.The west of Jilin Province is one of the most serious areas of land salinization in China.In recent years,the land salinization has led to the rapid deterioration of regional ecological environment,which directly restricts the regional ecological construction and economic development.At present,researches at home and abroad show that there are many methods to explore land salinization information and degree classification by remote sensing.However,due to the rapid change of land salinization itself,the extraction of salinization information is inaccurate and the classification accuracy is low.Moreover,the spatial heterogeneity of salinized soil leads to the uncertainty of degree change of salinized land in different regions.Therefore,on the basis of summarizing existing research methods,aiming at the problems of inaccuracy in the extraction of saline-alkali information and classification of saline-alkali degree in western Jilin,this paper takes the saline-alkali wasteland in western Jilin as the research object,utilizes medium and high resolution remote sensing image data,and conducts a classification study on the saline-alkali degree of land through multiple algorithms,providing reference for the comprehensive treatment of saline-alkali land in western Jilin.The research contents and conclusions of this paper are as follows:(1)Study on salinization degree classification of coarser algorithm for extracting alkali spot information based on high spatial resolution dataBased on the hyperspectral image data of GF-1 in western Jilin,the spectral response characteristics of saline-alkali land were analyzed,and the information of alkali spot in western Jilin with high spatial resolution was extracted by binarized image classification.The accuracy of classification results was tested by Kappa coefficient.The calculated K value was 0.69,indicating that alkali spot classification results had high consistency and high classification accuracy.Based on the analysis of the intrinsic scale of the ground objects in the saline wasteland in western Jilin,the maximum scale of alkali spot information coarsing algorithm was determined.On the basis of extracting the pixel of alkali spot by binary classification,the coarsing algorithm calculated the pixel rate of alkali spot and realized the grading mapping of alkali spot information coarsing algorithm.The grading results showed that the spatial distribution of different degrees of salinization information in different regions of western Jilin was quite different.The salinization degree is mainly mild salinization,and the land with severe and extremely severe salinization is mainly concentrated in the surrounding area of individual farmland,sandy land and water body.(2)Study on salinization degree classification based on mixed pixel decomposition of remote sensing data with medium spatial resolutionOn the basis of Sentinel-2 medium resolution multi-spectral image data,the end elements of caustic soda spot and pure grassland in the saline-alkali wasteland were determined according to the delimit space range of the saline-alkali wasteland.The linear spectral mixing model and the fully constrained least square method were used to decompose the mixed pixel elements of caustic spot in the saline-alkali wasteland in western Jilin.Salinization degree classification mapping of mixed pixel decomposition based on pixel base spot rate.The classification results show that the salinized pixels in the west of Jilin account for 60.37%of the total pixels,which is much larger than 43.84%in the classification result of the coarsening algorithm.The reason should be the difference of spatial resolution of remote sensing data sources.The salinization degree of the study area is still mainly mild salinization,and the moderate salinization area is more than the grading result of the coarsening algorithm.The land area of severe and extremely severe salinization is small and mainly concentrated in a few areas.(3)Construction of salinization degree classification model of saline-alkali wasteland in western Jilin combined with medium-high resolution remote sensing dataBased on the mapping results of alkali patch coarsing algorithm with high spatial resolution,the statistical model of salinization degree of barren grassland was constructed by using the regression model of unitary linear function model,quadratic polynomial function model,cubic polynomial function model,power function model,logarithmic function model and exponential function model.Among all the models constructed,It was found that the quadratic polynomial function model based on the salinity index SI had the best fitting effect on the salinity degree in western Jilin Province.The correlation coefficient(R~2)was 0.754,the determination coefficient(R~2)between the estimated value of alkali spot rate and the measured value was 0.85,and the root mean square error(RMSE)was 0.49,indicating that the model had high accuracy.It can be selected as the optimal model for the saline-alkali degree classification of saline wasteland in western Jilin,so as to realize the large-scale monitoring and classification mapping of the saline-alkali degree of saline wasteland.(4)Classification and optimization of saline-alkali wasteland in western Jilin by multiple methodsBased on the PH values of the ground measured samples,the classification mapping of alkali spot coarsing algorithm,the classification mapping of mixed pixel decomposition and the classification mapping accuracy evaluation of remote sensing model inversion were carried out.The optimal classification mapping method of alkali spot coarsing algorithm based on high spatial resolution data was found to be the salinity degree classification mapping of saline wasteland in western Jilin.The cartographic accuracy of this method is up to 70.2%,which can be used for further study,so as to realize the popularization and application of classification cartography of saline-alkali wasteland in other areas of Songnen plain.