International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
  
crop. IKONOS multi-spectral data of 30^ April 2001 shows the 
layout of the farm and the study field (Figure 1). 
  
Figure 1. IKONOS image of 30" April, 2001 showing the 
CPRS farm (# is the field where soil variability was 
studied) 
3. METHODOLOGY 
3.1 Remote Sensing Data 
IKONOS multi-spectral data of 30" April 2001 was acquired 
and during the period, when the field was fallow. The IKONOS 
satellite (Space Imaging, Thronton, CO, USA) was launched 
into low earth sun-synchronous orbit in September 1999. 
IKONOS provides images in panchromatic mode (0.45-0.90 
um), with Im spatial resolution and multispectral mode with 4 
m spatial resolution (table 1). The data is available in 8-bit or 
with full dynamic range 11-bit radiometric resolution. For the 
present study 11 bit data was used. The data is delivered in 
georeferenced format with UTM projection and WGS84 datum. 
Table 1. IKONOS spectral band characteristics 
  
  
  
  
  
  
  
  
  
  
  
Band Band Bandwidth Calibration 
Centre (nm) Coefficient 
(nm) [mW/(cm^*sr*DN) 
] 
Blue 480.3 71.3 728 
(MS-1) 
Green 550.7 88.6 727 
(MS-2) 
Red 664.8 65.8 949 
(MS-3) 
VNIR 805.0 95.4 843 
(MS-4) 
© For 11 bit products (post 22/02/2001) 
(Source: 
http://www.spaceimaging.com/products/ikonos/spectral.html) 
3.2 Soil Parameters 
Thirty-five soil samples (surface soil) were collected from the 
field at regular intervals. The samples were analyzed for soil 
organic matter (%), available nitrogen (ppm), available 
phosphorus (ppm), available potassium (ppm) and soil texture 
(sand, silt and clay percentage). The methodology for 
determining soil chemical (fertility) parameters is given in 
Table 2. Soil texture was determined using International Pipette 
method. As soil texture analysis a very time consuming process 
only twelve samples (every third sample) were analyzed. 
Table 2. Methodology for determining soil chemical parameters 
  
  
  
  
  
  
  
  
  
Parameter Method Reference 
Organic Matter Chromic acid Walkley and 
titration Black (1934) 
Available N Alkaline Subbiah and 
Permanganate Asija (1956) 
Extractable 
Method 
Available P Sodium Olsen et al. 
Bicarbonate (1954) 
Extr. Method 
Available K Ammonium Muhr et al. 
(Exch.+ acetate Extr. (1965) 
Soluble) Method 
  
3.3 Computation of Spectral Indices 
Various soil-related spectral indices were calculated from 
IKONOS MS data, after converting the digital numbers into 
radiance values. Those indices included soil related indices 
such as, Brightness Index (BNI), Hue Index (HI), Saturation 
Index (SI), Coloration Index (CI) and Redness Index (RI). The 
method for estimating these indices is presented table 3. Apart 
from these three principal components (PCI, PC2 and PC3) 
were also generated using principal component analysis (PCA). 
PCA has been used for dimensionality reduction of 
multispectral images in pattern recognition applications and for 
creating an optimal set of spectral information from a large 
number of bands. 
Table 3. Radiometric indices calculated using hyperspectral 
data for soil property study (adopted from Mathieu 
et al, 1998) 
  
  
  
Index Formula Index Property 
Brightness (R--G^-B^y3.0)? | Average reflectance 
Index, BI magnitude 
Saturation (R-B)/(R+B) Spectra Slope 
Index, SI 
Hue Index, HI 
Coloration 
Index, CI 
Redness Index, 
RI 
  
Primary colours 
Soil Colour 
(2*R-G-B)/(G-B) 
(R-G)(R+G) 
  
  
R^(B*G) Hematite Content 
  
  
  
  
3.4 Variability study of parameters 
Variability of soil and spectral parameters was analyzed by 
estimating coefficient of variation (CV) of the soil and the 
spectral parameters. The correlation analysis was carried out to 
study the relationship between soil and spectral parameters. 
Multiple regression models were generated, using stepwise 
regression technique, to estimate soil properties from RS data. 
The empirical models were generated only for those 
  
Internation 
pe ow 
parameters, 
equations W 
from RS dat 
The IKON! 
radiometric 
principal cor 
first princip 
variance in 
and first, se 
total varian 
components 
Table 4. The 
b 
Principal 
Compo- 
nent 
1 
  
Sun 
4.1 The Va 
variation of 
locations for 
soil has sar 
available N 
variability ai 
showed that, 
for available 
20.895). Am: 
PC3 (161.9% 
Table 5. Var 
Paramete 
O.M.(% 
Available N ( 
Available P ( 
Sand (% 
Silt (%) 
Clay (% 
[B 
MM 
SI