The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008 
1010 
1995 
119/31 
05.15 
7 
2001 
119/31 
08.11 
7 
Table 1. Faku TM data 
In addition, Faku county’s relief map of 1:50000 in 1975, 
investigation report of land use in 1992, population statistic data 
from 1995 to 2001 and other text data were collected. 
3.2 Remote sensing data preprocessing 
Data preprocessing mainly refers to do data correction, noise 
removed, cut, and so on before the official use [2 l The chosen 
data has no spots or various types of noise, So do data 
preprocessing including putting the text data in database, 
geometric correcting and radiometric correction for the remote 
sensing images. 
3.3 Determination of NDVI threshold 
Vegetation index reflects the approximation relation between 
the spectral response and vegetation cover. The essential 
characteristic of desertification is the lower the productivity of 
land. Changes in vegetation index can reflect the changing 
process of land productivity. Therefore vegetation index can be 
used as desertification monitoring indicators to monitor land 
desertification and dynamic changes. NDVI is the abbreviation 
for Normalized Difference Vegetation Index. It was confirmed 
according to strong absorptions of chlorophyll in 0.69pm. Use 
ratio between infrared and near infrared and linear combination 
to express vegetation information. The Vegetation index can 
reflect a series of physics quantity of the plants, such as index 
of leafs area, plant overlays degree, amount of living creature 
and plant category etc [3 ' 4] . The index of vegetation contain 
many different expression methods, among them NDVI is 
applied in the fixed amount research of plant overlays. 
Calculation formula of NDVI is: 
For the image of TM: 
ND VI = (p4 - p3) !(pA + p3) I 
p3 = TM images the three-band visible light 
Determining the threshold of NDVI is the first step in extraction 
of land desertification. Then we can use the threshold to 
separate vegetation information and non-vegetation information 
from land cover. 
Through comparative analysis of the NDVI images and the 
research area, select NDVI = 0.05 as the threshold, which can 
separate vegetation information and non-vegetation information 
from the research land cover, that is: if NDVI> 0.05, the area 
was considered as vegetation; Instead, considered as non 
vegetation information. The NDVI image of research area in 
2001 is below (Figure 1). 
Figure 1. NDVI image of research area in 2001 
3.4 Desertification index system of research area 
Through the sufficient district investigate and practice research, 
divided the extent of desertification in the research area into 
three degrees: slight desertification, middle desertification, 
severe desertification. Specific quantitative indicators are 
shown below (Table 2). 
3.5 Result of the research area’s classification 
Do remote sensing classification based on the index system in 
the research area. The result is below(Figure 2.). 
Where 
p4 
vegetation coverage 
NDVI (Normalozed Difference 
Vegetation Index) 
PCA (3,2,1) synthesis features 
of color images 
slight desertification 
<15% 
<0. 14 
shallow green, distributed in the 
stope edge, the bands on both 
sides of the river 
middle desertification 
<10% 
<0. 10 
shallow pink, distributed 
around severe desertification 
severe desertification 
<5% 
<0. 07 
red or blue, distributed in the 
waste dump, mining pits and 
the surrounding settlements 
Table 2. Extent of desertification hierarchical list in Faku county