International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B-YF. Istanbul 2004 
  
The study area takes 600mm rainfall in average of a year. In this 
production farm, mostly wheat is the main product and has 
approximately 3000-ton capacity in a year. In addition, 
sunflower, corn and canola are the other agricultural products 
that are aggregated. 
In this study, three multi temporal Landsat-TM and one SPOT- 
XS data sets were used to analyse the vegetation biomass 
changes over time. The characteristics of satellite data used are 
shown in the Table 1. 
  
  
  
  
  
  
  
  
  
  
  
Spectral Spatial 
Satellite Date resolution resolution 
(um) (m) 
Band 1 | 0,45 - 0,52 30 
11.05.1987! Band2 | 0,52- 0,60 30 
Band3 | 0,63 - 0,69 30 
27.05.1995 | Band4 | 0,76 - 0,90 30 
Landsat Band5 | 1,55-0,75 30 
TM 
Band6 | 10,4- 12,5 120 
07.06.2000 
Band 7 | 2,08 - 2,35 30 
Band I | 0.50-0.59 20 
Sot 12.05.2003 | Band 2 | 0.61-0.68 20 
Band 3 | 0.79-0.89 20 
  
  
  
  
  
  
  
Table 1. The characteristics of satellite data used. 
3. METHODOLOGY 
3.1 Vegetation Indexes 
In this study, five different types of vegetation indexes, which 
quantify the concentrations of green leaf vegetation around the 
globe, were used for biomass analysis. These indexes depend on 
the reflectance of vegetation, which is very different in near 
infrared and red bands. Healthy vegetation should absorb the 
visible light and reflects most of the near infrared light, on the 
other hand unhealthy vegetation reflects more visible light and 
less near infrared light. The reflection on visible band is related 
with the pigments in the leaves of plants but in the near infrared, 
it depends on the cell structure. 
Taking the ratio of near infrared band and red band is the 
simplest vegetation index. Hence, it is called Simple Ratio (SR) 
or Ratio Vegetation Index (RVI). SR indicates the amount of 
vegetation. In the resultant SR image, high values, such as more 
than 20, show for dense vegetation and low values, which are 
around the value of 1, show for soil, ice and water. However, it 
doesn't give information related with topography. It only 
transmits the spectral information; therefore this also gives an 
opportunity of having uniform spectral classes after 
classification. 
Another simple vegetation index is the Difference Vegetation 
Index (DVI) which is also sensitive to the amount of the 
vegetation. Mathematically, it is in the form of (near infrared 
band) — (red band). DVI has the ability to distinguish the soil 
and vegetation but not in shady areas. Hence, DVI doesn't give 
proper information when the reflected wavelengths are being 
affected due to topography, atmosphere or shadows. 
The more common and known one is the Normalised Difference 
Vegetation index (NDVI). The algorithm of NDVI is (near 
infrared band- red band)/(near infrared band + red band). 
Resulted values change between -1 and +1 regarding to the 
vegetated area. Such as, if the result is 0,1 or below, it 
corresponds to an area of rocks; if it is between 0.2 and 0.3, it 
indicates an area of shrubs or grasslands; if it is between 0.6 and 
0.8 it corresponds to an area of tropical rainforests. 
Transformed Normalised Difference Vegetation index (TNDVI) 
is the square root of the NDVI. It has higher coefficient of 
determination for the same variable and this is the difference 
between TNDVI and NDVI. The formula of TNDVI has always 
positive values and the variances of the ratio are proportional to 
mean values. TNDVI indicates a relation between the amount of 
green biomass that is found in a pixel. (Senseman et.al. 1996) 
Perpendicular vegetation index (PVI) is one of the complex 
indices that also including soil emissivity factor. It is based on 
the linear relationship of red and near infrared reflectance from 
bare soils. This is called the soil line (Figure 2). PVI is the 
perpendicular distance from the soil line and it is linearly 
related to the vegetation cover (Sunar and Taberner 1995). PVI 
uses Gram-Schmidt orthogonalization to figure out the 
greenness line, which is perpendicular to the soil line and passes 
through the %100 vegetation cover points. PVI is effective in 
detecting dry and green vegetation. This is caused by the 
sensation of red and near infrared combination to the iron oxide 
absorption that is in many soils. Mainly PVI indicates the 
vegetative cover, independent from the soil effects. 
MEASURED 
REFLECTANCES 
NIR 
REFLECTANCE 
1." SOIL LINE 
  
REFLECTANCE 
RED 
Figure 2. Perpendicular Vegetation Index. 
3.2 Geometric Correction 
To detect the changes in vegetation biomass all images used 
must be registered to each other. The 1993 Landsat image was 
taken as the base image for registering. 
Ten GCPs for each year, which were well distributed through 
the images, were chosen in registration. The number of GCPs 
and rms errors were outlined in Table 2. Spot XS image (2003) 
were resembled to 30 m to be analysed together with the other 
Landsat images. All the registered images were taken as a 
multitemporal dataset having 455 x 547 pixels. 
  
  
  
  
  
Base image | Slave image # of GCPs rms error 
1987 10 0.5716 
1993 2000 10 0.5474 
2003 10 0.4373 
  
  
  
  
  
182 
Table 2. Number of GCPs used and rms errors.