To achieve this objective the following steps has been done:

    From IR-IC-LISS-III images we generated False Colour Composite (FCC) using (green-red-NIR) bands, which shows different  areas represent Agriculture and Urban areas and geomorphological features like: Prominent meander cut-off, ox-bow lake,  successive paleo channels. Geomorphological studies has been carried out, which indicate that the Ganga river first migrated is an eastern direction and then started its west ward movement.  The images clearly show the presence of two distinct geomorphic zones, in the Western Zone elder layers of flood plain characterised by modified or overlapping paleo channels.  The eastern part characterised by development of meander scars, sloughs, and lakes.  In this zone younger deposition then western part.  One more thing has been taken out the discontinuity of geomorphological features like paleo-channels discontinuity and change in their trending direction.  That indicates some of these paleo-channel exposed to stress, which accompanied with fault occurring. Vegetation index (NDVI) has been calculated for the same area using band 4 and band 3 (NDVI= B4 - B3 / B4 + B3), NDVI values show also two different area, western zone highly vegetated of the other part shows negative values for NDVI due the presence of water, and poorly vegetated.  After calculating vegetation index classified image has been produced depends on NDVI values. To study edge detection from an image the following filters has been applied.

    Sobel operator which look for edges in both the horizontal and vertical directions and then combine this information into a single metric.  Sobel edge operator applied on red and NIR bands, band 3 give better results then band 4, and shows completely the main paleo-channel of river before neck cutoff occurring, and also show two parts the western part more edges due to presence of paleo channels and the eastern part has less edge or no edge at all.  This contrast due to presence of surface water. Laplacian edge operator has been tested here, but it did not give any significant results. High pass filter :  This filter will remove the low frequency components and show only high detail. Applying this filter on band 3 give the ability to define the location of the fault on an image.  (High pass = original image - mean filter applied) on the same image, this is the equation has been used in this study.

  To improve the spatial resolution and consequently to improve visual interpretation merging technique has been applied using four methods.

    IHS method consists of transformation of Low Spatial resolution data set from RGB to HIS co-ordinate system and then the
intensity component is replaced by panchromatic image before carrying out the inverse transformation of HIS to RGB space.

    P+XS Method :  This method based on the assumption that the panchromatic band is the sum of bands Xs1 and Xs2. This method has been modified to be fit with IRS -1C LISS III data.

    HPF in this method high pass filter is applied to high spatial resolution data, and mean filter applied on an low resolution data and the results added pixel by pixel together.

    Price method (Price, 1987) each new  band is the result of the combination of the original spectral band and an estimation of this band in terms of high resolutio band.  For the spectral bands closely correlated with the high resolution image, the estimation is obtained by linear regression, whereas for weakly correlated bands a look up table is prepared to obtain the high resolution estimate.

    Applying the above methods, HPF gives better results based on number of zero pixel values compared to other methods are used. Price techniques gives second better results.  This is because of using different time for multi-spectral and Pan bands.   Only HPF and price method has removed cloud cover from merged mages. Of course merging techniques has improved visual interpretation. Before merging multi-spectral images show that the western part of main paleo-channel is too steep while after merging clearly show no steepness in that part.  It is clearly, after merging this edge comes due to presence of road in the same area, not because of fault.

  Applying Otsu thresholding techniques to remove noise from an image and to see the effect of merging on edge extraction the Otsu thresholding techniques has been applied on an imagesbefore merging and after merging.  The ideal of Otsu Thresholdingtechniques is to select thresholding value from histogram has adeep and sharp valley between two peaks representing objects and background respectively, so that the threshold can be chosen at the bottom of this valley.  After knowing the thresholding value, we generated binary image which contains two pixel values "0" and "255".  Pixels with grey value "255" indicate an edge pixel. Thresholding will remove unwanted edges from an image. Applying this techniques has given good results using images before merging better then applying it on images after merging red band gives better edge enhancement before and after merging compared to NIR band.  From binary image we can easily recognize between two different object and digitize the fault which lying there.

    Resistivity Survey: To prove and correlate the results of remote sensing with resistivity variation, using of resistivity meter , "14" sounding point carried out in the field around the expected area. Resistivity data shows difference in Apparent resistivity in both sides, especially surface layer  which give good contrast between two sides and also undisturbed layer in both sides.  The apparent resistivity in western side show anomaly in resistivity readings that indicate the presence of disturbed layers due to one tectonicevents.  The other part eastern part show uniform changes in resistivity which indicate this part contains younger layer deposited in quiet conditions.  In the field many things has been noticed like disturbance in ground water taste and quality around the location of fault, and change in soil characteristics while moving from South to North in that area.

    From both remote sensing and resistivity variations the
following results has been obtained :

While Ganga river was migrating in the eastern direction the main major tectonic factor happened.  That was a fault in the basement (NE, SW) Normal fault.  This fault cause uplift to the western part and merges the eastern part.  This fault associated with another fault located out of my area produced a mini garben, and this very good place for new sediment deposition.  The continuity of clay and silt and fine particles deposition in quiet conditions (like lake conditions) associated with climatological conditions help to produce this thickness of soil in the eastern part.  At that time channel migration through neck cutoff happened. It is clearly that the depth of basement is more in the eastern part compared to western part.  Then remote sensing data associated with geophysical data give a good idea about the tectonic in the basement.