Oasis is an important component of desert ecosystem. This paper employs Landsat Thematic Mapper (TM)
multi-spectral data to extract fractional vegetation cover of oasis in Tarim Basin with four methods. The mixture pixel
decomposition model based on normalized difference vegetation index(NDVI) is firstly used to estimate fractional
vegetation cover(FVC). The results indicated that the method is mainly underestimating the FVC at the low FVC area
and overestimating the FVC at high FVC area. Thereafter, a stepwise regression model between 15 Vegetation Indices
(VIs) and measured FVC data and a log-linear model have been established through the relation analysis of FVC and
NDVI. Trials of these two models showed that they are mainly overestimating the FVC. Finally, a dimidiate fractional
cover model was proposed, which is composed of two linear functions. When the NDVI is less than 0.3, the linear
function is formed by stress related vegetation index (STVI1) and normal differential water index (NDWI) (R2, 0.764)
while the NDVI is greater than 0.3, the linear function is composed of NDVI and perpendicular vegetation index (PVI)
(R2, 0.801). The validation of the dimidiate fractional cover model has been tested with the measured data. In the
optimal case, the mean error is 0.002 and the RMSE is 0.051, demonstrating that the model can be used in estimating
fractional vegetation cover of oasis in Tarim Basin.
The tasseled cap transformation of remote sensing data has been widely used in environment, agriculture, forest and ecology. Tasseled cap transformation coefficients matrix of HJ multi-spectrum data has been established through Givens rotation matrix to rotate principal component transform vector to whiteness, greenness and blueness direction of ground object basing on 24 scenes year-round HJ multispectral remote sensing data. The whiteness component enhances the brightness difference of ground object, and the greenness component preserves more detailed information of vegetation change while enhances the vegetation characteristic, and the blueness component significantly enhances factory with blue plastic house roof around the town and also can enhance brightness of water. Tasseled cap transformation coefficients matrix of HJ will enhance the application effect of HJ multispectral remote sensing data in their application fields.