Land cover information extraction based on ALOS image by object-oriented method

According to the high resolution remote sensing image data， this paper used the object-oriented classification method to extract the land cover information of Nanjing city. Because the classification unit was not single pixel but image objects， which integrated the spectral information， texture feature， topological relationship and thematic information， the overall accuracy of classification enhanced by 18.1% compared to conventional method. The object- oriented method also overcame the pepper salt phenomenon which occurred in conventional method， and reduced the negative effects of foreign bodies with spectrum and different object with the same spectra characteristics. For the land cover information extraction and updating using high resolution remote sensing image， this method had practicable and generalization performance.