ESAEducationHomeWeather and ClimateGlobal ChangeNatural Disasters
   
Earthquakes
IntroductionIzmit earthquake
Floods
Thessaloniki (Greece)
Hurricanes
Storm surgesHurricane Katrina
Volcanoes
IntroductionAndean Volcanic BeltEtnaNyiragongo and Nyamuragira
 
 
 
 
 
printer friendly page
Exercise 3: Change detection – Multi-temporal analysis
 
RGB false colour composite image
RGB false colour composite image
 
In this exercise you will perform change detection by creating false colour composites, using the images acquired before and after the flood event. All changes will appear in different colours; by assigning the post-flood image to the red and green channels and a pre-flood image to the blue, flooded areas appear as blue (because in those areas, blue will have a much higher value than green and red), while wet soil will be yellowish (because green and red will have high values, much higher than blue). Unchanged features (e.g. buildings in villages) will appear in different levels of grey.
 
Open 05102004_f.tif, 25102005_f.tif, 10102006_f.tif and pre-flood_mean.tif in one dataset.

Create an RGB false colour composite, by assigning:

  • 10102006_f.tif image to the Red and Green bands and
  • 25102005_f.tif image to the Blue band

The resulting colours show the changes between 25102005 (pre-flood period) and 10102006 (post-flood period). Repeat the same process by changing only the Blue band, and projecting the 05102004_f.tif image instead. The resulting colours display the changes between 05102004 (pre-flood period) and 10102006 (post-flood period).  
 
Now create one more RGB composite by using the post-flood image and the arithmetic mean of the pre-flood images that you calculated in the previous section, using the following combination:

R band: 10102006, G band: 10102006, B band: pre-flood_mean

Save the resulting image as RGB.tif.
 
 

1. Examine the RGB.tif image and describe what you see.

2. Can you identify the two main changes related to the floods (flooded areas and wet areas)?

3. If you wanted the flooded areas to appear in red colour, which band combination would you choose? What would then be the colour of the wet areas? Try to experiment!

4. Determine the land cover types that are covered by the two kinds of changes. In order to do that, use the information that you extracted by inspecting the pre-flood images during the first exercise.

5. Measure the size of the most prominent flooded areas in km2.


 
 
 


Flash floods in Thessaloniki
IntroductionBackground
Exercises
Exercise 1: Inspecting the imagesExercise 2: Change detection – band mathExercise 4: GIS
Eduspace - Software
LEOWorks 4 (Linux)LEOWorks 4 (Windows)LEOWorks 4 (MacOS)
Eduspace - Download
Thessaloniki.zip
 
 
 
   Copyright 2000 - 2015 © European Space Agency. All rights reserved.