taluk level using satellite remote sensing data base on sound 
statistical and image processing techniques. 
METHODOLOGY 
The study area consisted of a cluster of six villages, viz., Melur, 
Mallur, Muttur, Kachahalli, Ganganahalli and 
Patalammanahalli, under Melur TSC, Sidlaghatta taluk. The 
total geographical area is about 11.22 sq. km. 
Data Used: 
Satellite data: 
The IRS-1C-LISS-IH and PAN digital data of path 100 row 64 
acquired in 3 different dates of LISS-III data (in 24 days 
intervals) has been employed, starting from 03.03.1997 to 
20.04.1997, for mulberry garden condition assessment . 
Ancillary data: 
Survey of India topographical maps on 1 : 50,000 scale were 
used along with village / cadastral maps obtained from the 
Department of Survey, Settlement and Land records, GOK. 
Ground truth data: 
“ The ground truth data collection has an important bearing on 
most of the remote sensing applications. The phonological 
observations on mulberry crop such as growth stage, vigor, 
ground cover percent, soil type, etc., were collected in detail on 
survey number wise ( Medhavy, et al., 1993; Dutta, et al., 1992 
) Ground truth data collection in all the six villages, survey 
number wise was synchronised with the satellite passes. 
Digital analysis: 
The digital data analysis was carried out at RRSSC, Bangalore 
on IBM RISC 6000 system using EASI / PACE software and 
“CAPE MANAGER’ software developed by RRSSC, Nagpur 
and SAC, Ahmedabad. The methodology followed in the 
present study is described briefly below (Anonymous, 1990b; 
Bauer, 1985). : 
Extraction of data: 
The administrative boundary of Sidlaghatta taluk was digitised. 
A merged product is generated using IRS-1C-LISS-III and 
PAN digital data. Initially, georeferencing of the satellite data, 
which was carried out for acreage estimation has taken as 
georeferencing of the satellite data covering an area of a cluster 
of six villages of Sidlaghatta taluk. For image-to-map 
rectification a total of 12 GCPs were located in the image with 
reference to 1 : 50,000 scale toposheet. Since the image size 
was small, first order polynomial geometric transformation was 
found quite adequate between map-to-image co-ordinates of 
GCPs. The accuracy achieved through this simple linear 
transformation was at  subpixel level in both X and Y 
direction. Sub-images covering a cluster of six villages were 
also extracted from data of remaining two dates. Then image to 
image registration was done with respect to the rectified master 
image ( Anonymous, 1986; Geeta Vardhan, et al, 1992 ). 
Radiometric resampling of pixels using nearest neighbor 
techniques was applied as this enables pixel by pixel 
comparison of multi-temporal change during the growth cycle. 
The image of multi-temporal data of part of Sidlaghatta taluk 
was generated. It consists of three dates with four bands each 
was extracted separately and stored for further analysis. The 
LISS-III data was digitally resampled and registered with edge 
enhanced PAN data using image to image transformation 
model. Hue Intensity Saturation (HIS) transformation was used 
to merge the information contents of both the data sets. 
IAPRS & SIS, Vol.34, Part 7, “Resource and Environmental Monitoring”, Hyderabad, India, 2002 
Cadastral map was overlayed on merged data on 
approximately 1 : 12,500 scale. The methodology comprises 
satellite data processing, generation of raster cadastral images, 
registration of cadastral images, generation and overlay of 
cadastral vectors over the digitally classified images. These 
hybrid images depict field boundaries, cart tracks, small nullahs 
besides the settlements, tanks and other cultural features like 
roads and canals. Most of the survey / parcel boundaries on the 
cadastral maps match well with these features. This is an 
advantage in identifying GCPs for developing transformation 
model for image to image tie down ( Geeta Vardhan, et al., 
1992 ). 
Generation of raster cadastral images: 
The cadastral map was scanned and corresponding raster 
images were generated. The thickness of parcel boundaries 
were thinned using the Kernel based software. 
Registration of cadastral images: 
The cadastral images were registered with high resolution edge 
enhanced colour composite using large number of GCPs. An 
accuracy of less than half a pixel of standard and residual errors 
were achieved while registration of cadastral images. Due to the 
resampling of the radiometric values the parcel boundaries were 
found to discontinuous lines and to avoid the problems of 
artifacts and null points, a special software developed at the 
RRSSC, Bangalore was used for the cadastral maps 
registration. The resultant output depicts continuous lines of 
the parcel boundaries ( Rao et al., 1996 ). 
Generation and overlay of cadastral vectors: 
The registered parcel boundaries were overlaid on satellite data 
in the form of raster bit maps or vectors. Vectors overlay is 
preferred to raster overlay because of its unique line thickness 
consistent with the reduction / enlargement of the raster images 
and optimal utilisation of the disk space. Vector boundaries 
were generated either from conversion of raster image or by 
digitisation of raster boundaries ( Radhakrishna, et al., 1996 ). 
Mulberry crop condition assessment procedure: 
The study area is extracted using the mask, covering six 
villages, for digital analysis of three dates and subjected to 
further processing to derive information on mulberry growth 
conditions. Flow diagram illustrating the methodology for crop 
condition assessment is shown in Fig. 3H. 
Generation of vegetation index: 
Vegetation Index is a ratio based on reflectance in the Red and 
NIR bands of the electromagnetic spectrum. Vegetation Index 
(VI) images were derived from empirical transformations of 
spectral responses in different spectral bands. Normalised 
Difference Vegetation Index (NDVI) used for monitoring the 
growth conditions of mulberry in the present study was 
generated by using the equation given below: 
NDVI=IR-R/IR+R 
Where, IR - reflectance in the infrared band ; R - reflectance in 
the visible red band and NDVI - normalised difference 
vegetation index. The NDVI images were generated for all the 
three satellite data sets. 
Generation of crop growth profile: 
Digital data of IRS -1C-LISS-III of 3 dates in a crop season 
were used to generate mulberry crop growth profile. The 20th 
April, 1997 image registration was done with respect to the 
rectified master image. 
    
zh 
( 
7 FO d3 OO SZ 
fc 
a 
te