| | Location of the study area, Khumbu Himal - Part 4
Working with Landsat images
Open the image 'khumbu_321.tif' and, with the help of the map you used before, zoom into the Imja Tso region.
|  | Area around the Imja Tso in false colour | |
The first three bands of the Landsat satellite are within the visible spectrum of solar radiation. This means that the band combination 3,2,1 shows the image of the Earth’s surface the same way human beings would see it with their eyes.
In the lower left frame of the viewer, several numbers are displayed. In the first brackets, the position of the cursor within the image is shown. The other numbers show us the value of the file pixel of the position of the cursor. As this image is displayed in 8 bits, the range of the value is between 0 and 255 (2*2*2*2*2*2*2*2 = 256). The higher the value, the brighter the pixel displayed.
For this exercise, take a sheet of paper and a pencil, or construct a table like the one below. Take the image from the previous page as a reference for land cover localisation. Look out for the boxes with numbers by referring to the table below. First, place the cursor on the lake, which is easy to identify on the glacier. Write down the cursor position and the values of the pixels for each spectral band (red, green, and blue). Always look for y typical value, not extreme high or low.
Move the cursor around over different kinds of land cover classes (water, moraine, vegetation, snow in sun and shadow, ice, gravel…) and write down their pixel values in a table like the one below. Keep the cursor position identical (as close as possible) for all bands for the specific land cover.
For the 'khumb_321' combination, read the pixel value for all bands. B is the blue band, G is the green band, and R is the red band. Note that on the image window the order is different. In fact it is referred to as 'RGB'. For the other combinations, note only the first one (red). Always read the values of the land cover in the same (as close as possible) corresponding cursor position location.
As a next step, open the image 'khumbu_432.tif' which you created before, and analyse the same land cover classes. Try to get the same pixel. Just enter the value for 'R' into the table for each land cover.
1. When you compare the pixel values from the same class in the other image, what changes?
2. When you look at soil or vegetation in the image, on which image are they better separated, on the 321 or 432 image? Explain the advantage of using an infrared band.
As a further step, we will open the image 'khumbu_543.tif'. Here we have not only one band in infrared, but two. Just enter the value for 'R' into the table for each land cover.
3. Guess what changes will appear in the image.
4. Again, compare and write down the values of the pixels of the classes in the table.
5. When you look at the different images, which classes are best displayed, and in which image?
As a last image, open the file 'khumbu_752.tif'. Just enter the value for R into the table for each land cover. Do the same steps for this image and compare it again with the others.
| |
