Results and Disscussions
1.1. Visual Interpretation of The Images
In order to detect the affected and non-affected tea patches the temporal images were visually interpreted. The false colour composite (FCC) for LANDSAT, LISS III and ASTER were generated at 4:3:2, 3:2:1 and 3:2:1 respectively. From the FCC the affected and non-affected tea patches could be distinctly separated from the other landuse/landcover classes. From the images (as shown in figure 13) it is quite clear that the healthy patches showed bright red appearance, moderately affected patches had reddish brown appearance while the diseased patches showed dark brown appearance.
The landcover classes were described visually as:
- Tea Patches: From the image the tea patches appears to be bright red in colour with smooth texture. The healthy tea patches showed bright red colour due to higher reflectance while the moderately affected patch showed reddish brown in colour. The diseased tea patch showed dark brown colour.
- Other Vegetation: Other vegetation like scrubs gives pink to dark red in colour with both smooth and coarse texture.
- The river gives dark blue reflectance with smooth texture that can be easily separated from the other classes.
- River Bed: The river bed also gives whitish appearance that could be easily distinguished from all the other landcover classes.
- Barren Land: Barren Land gives wheatish green appearance with smooth texture that could be well separated from the other vegetation types.
- Settlements: Settlements shows blackish green reflectance with coarse texture.
Figure 13: (Sonitpur District of Assam, LANDSAT ETM+ (December, 2001)
1. Tea Patches 2. Barren Land 3. River 4. Forest 5. River Bed 6. Settlements