The Study Of Land Use Classsification Based On Remote Sensing

With the progress of science and technology, remote sensing technology is getting mature and has been widely used, due to its own characteristics, in the classification of land use. The classification of land use is one of the fundamental jobs and the application of remote sensing technology in it can obtain objectively and accurately the information of land use by effectively using comprehensiveness and timeliness of remote sensing data. As different satellite remote sensing data have their own advantages and disadvantages, good applicability and application value can be achieved through combining multi-source remote sensing data and using their different spectral and temporal resolution in the classification of land use.In this paper, Gaochun County, south of Nanjing, is selected as the study area. The study, based on the combination of ETM+ remote sensing image data and MODIS remote sensing image data, with various index as the classification reference data, using various classification algorithms to compare and analyze, aims to discuss and study the classification of land use in Gaochun County. On the basis of the study, the following conclusions can be got:1. Combination of ETM+ remote sensing image data and MODIS remote sensing image data can make full use of MODIS data’s characteristic of multi-spectral and multi-temporal, and can utilize the characteristics of high resolution in Landsat ETM+ image data. At the same time using this method can improve the accuracy and feasibility in the aspect of selecting training samples and evaluating the accuracy of the results.2. Selection of classification feature will directly affect the accuracy of the classification. Single classification feature has been unable to meet the increasingly complicated demands of classification. The application of extracting multiple classification features with MODIS data to combine and optimize in the classification of land use can have a good effect and applicability.3. Traditional decision tree algorithm has been lagged behind because of its spectra characteristics. CRUISE algorithm becomes better in the way of space expression, arithmetic speed and classification accuracy. Using CRUISE algorithm with MODIS data’s characteristic of multi-spectral and multi-temporal to classify the land use type can improve the accuracy of the classification and better certain spectral characters.This paper, on the basis of the combination of multi-source remote sensing data, with various classification features and improved Classification algorithm, classifies the land use type in Gaochun County. The study, as a result, can make full use of the advantages of multi-source remote sensing data and get high accuracy and Kappa coefficient. Also, it well reflects the actual situation of the local land use and has good practicality and fine prospects for development.