Extraction of land use classification information of suburb of Nanjing city using ASTER image

Land resources are essential substance foundation for development of society and economy. Land resources can be classified and investigated by using remote sensing information. It is shown that high spectrum resolution images can better offer higher quality information source in observing the earth's surface in comparing with previous common remote sensing images. ASTER is an advanced multi-spectral sensor with high spatial resolution on the Terra satellite. It is a general-purpose instrument that covers a wide spectral region from the visible to the thermal infrared by 14 spectral bands (with high resolution from 15 m to 90 m) for monitoring natural resource and ecological environment changes, i.e., land use and land cover, short-term climate variability, natural disasters in the Yangtze River Valley. The landuse classification of suburb of Nanjing city was studied using ASTER image. In this paper, the multivariate statistical analysis was used to retrieve classification band, and then initial training samples were derived from unsupervised classification. After purifying pixels in region of interest time after time, status map about suburb of Nanjing city was classified by using supervised classification method. The result of supervised classification shows that the classification method combined with other image treatment methods can better retrieve landuse information from ASTER image and obtain high precision.