Accuracy Analysis Of Land Use And Land Cover Classification Based On Different Classification Methods

Remote sensing(RS)image classification is always a pivotal part of remote sensing study. How to improve the accuracy of RS interpretation is an urgent problem in RS application.Based on the practical application of Land use and land cover information extract, the Landsat 5 TM remote sensing imagine of Hongqinghe County of Yijinhuoluo Banner in Inner Mongolia Autonomous Region is taken as source data. After analying the spectral characteristics of TM imagine and selecting the optimal bands combination, the remote sensing imagine of this region is studied in classification by three different ways, which are Maximum likelihood classification, Texture analysis and Expert classification method of decision tree. Also the classification accuracies three different ways are compared, according to the comparison results that the classification accuracy is analyzed to every Land use and land cover types.According to the analysis and evaluation to the evaluation index of classification accuracy, the conclusion shows that the overall accuracy of three ways are all 80% and more which can satisfy the demand of application. That using the texture information of imagine in classification can improve the accuracy. Expert Classification method of decision tree is of higher overall accuracy in classification, but is of a sort in the classification of same types of land use and land cover.Compared with traditional Maximum likelihood classification method, the classification accuracy of Decision tree classification method and Texture analysis method can improve classification accuracy and effects of Land use and land cover information, but their principles and calculations are more complex. And every ways had its advantage and disadvantage.