Land cover of oases and forest in XinJiang, China retrieved from ASTER data

Abstract

ASTER aboard NASA’s satellite Terra is a high-resolution multispectral radiometer of 14 bands. The spatial resolution is 15 m in VNIR, 30 m in SWIR and 90 m in TIR spectra, respectively. With the data observed with ASTER, the land cover classification is produced for the Tarim Diversifolious Poplar Protection Area along the Tarim River in the northern Tarim Basin (Taklamakan Desert) in XinJiang, China. The classification of the vegetation (plants) in the arid and semiarid regions using remote-sensing technology is very difficult. Because the cause has low vegetable cover density and the influence of reflection from background soil is large. ASTER data are effective in studying the spectrum characteristics of land cover in arid and semiarid regions. The sensor has several bands in the shortwave infrared wavelength region that is designed for exploration of earth resources and study of the arid and semiarid region natural environment. However, we are not clear combination of which band is the most effective in research of the arid region like the Taklamakan desert in the data of 14 bands of ASTER. The optimum index factor (OIF), based on total variance within bands and correlation coefficient between bands, is a statistical approach to rank all possible three-band combinations. In the process of analyzing the data, the pixel sizes of all the data are converted (layer stacking and re-sampling) into consistent same size of 15 m. The three-band composite with the largest OIF value will have most information (as measured by variance) with the least amount of duplication (as measured by correlation). We used the OIF technique to rank all three-band combinations of ASTER original 14-band data over Tarim River Poplar Protection Area. Our study indicates that RGB color overlay using atmospheric corrected ASTER original bands 2, 3 (VNIR), and 6 (SWIR) has the highest OIF. When NDVI is considered as one ASTER band, highest OIF will have by carrying out bands 3 (VNIR), 4 (SWIR), and NDVI. In this study, we used highest OIF (bands 3, 4, and NDVI) succeeded in extraction of Tarim River Poplar Forest.