Research On Applicability Of Spectral Change Vector Analysis For Land Use Monitoring In Eastern Inner Mongolia

Spectral Change Vector Analysis method can not only use multiple bands for data processing,make full use of the spectral information of the ground objects,and get all the change information of each pixel in the image,but also can avoid the error caused by multi-classification,the application of change detection and change type determination in land use dynamic monitoring has been paid more and more attention.This thesis,took Zhagasitai town,Ar Horqin Banner,Inner Mongolia as the experimental object and studied the dynamic monitoring of regional land use by using the method of Spectral Change Vector Analysis,so as to analyze the applicability of this method in the eastern part of Inner Mongolia.It provides a reference for the application of Spectral Change Vector Analysis in land use dynamic monitoring,and has important experimental value in improving the accuracy and efficiency of land use dynamic monitoring.In this thesis,Landsat remote sensing images from July to September from 1985 to 2018 were used to obtain the data sets of apparent reflectance and surface reflectance of the study area by preprocessing the 8 phase remote sensing images through apparent reflectance inversion,surface reflectance inversion,mosaic,etc.,at a time interval of 5 years.Based on the Spectral Change Vector Analysis method,firstly,the change thresholds were determined by the image difference method,and the land use change pixels were extracted.Then the land use change types were determined by the parallelepiped classification method in the cosine space of the spectral change vector direction in the range of detected pixels,thereupon the land use monitoring in the study area was achieved.Finally,the accuracies of dynamic monitoring of land use in the study area by using Spectral Change Vector Analysis method based on apparent and surface reflectance data were evaluated by confusion matrix method using land use interpretation results from 8 periods remotely sensed images,then the applicability of the Spectral Change Vector Analysis method in the study area,even in the eastern part of Inner Mongolia was analyzed.The results show that the average overall classification accuracy of land use type change information extracted from the apparent reflectance data set is 87.51%,and the average Kappa coefficient is 44.47%,the average total classification accuracy of land use type change information extracted from land surface reflectance data set is 87.74%,and the average Kappa coefficient is 48.92%.It is found that the results of land use dynamic monitoring based on land surface reflectance data set are better and slightly accurate than the land use dynamic monitoring results which based on apparent reflectance data set.The research shows that the precision of the results can meet the demand of fast dynamic monitoring,and the method can be used to monitor the land use in the east of Inner Mongolia.