 |  |  |  |  |
|  | |  | |
 |
Earthquakes IntroductionIzmit earthquakeFloods Thessaloniki (Greece)Hurricanes Storm surgesHurricane KatrinaVolcanoes IntroductionAndean Volcanic BeltEtnaNyiragongo and Nyamuragira
|  |  |  |  | | Exercise 2: Lava Detection using Supervised Classification - continued
Training fields
- Choose an RGB band combination (e.g. 4-5-3) and stretch the image so that you can visually distinguish between different classes.
- Multivariate Analysis >Supervised Classification > Select Training Fields or press on the red TF icon. A new bar will open.
- Press the open icon in TF bar
- Open the file TF_2001_6cl.sav
- Inspect some of the training fields and look at the classes.
Training fields for Landsat 2001 (FCC 4-5-3), 6 classes To learn how to work with the GIS module, please refer to the tutorial on p. 92.
To draw some more training fields:
- Digitise an area.
- Finish the polygon with a right click on the mouse.
- In the training fields window name the polygon according to the class and then press enter.
- Save the result (the icon is in the TF bar).
Running the classification, assigning colours to classes
- Open the un-stretched (original) Landsat 2001 bands number 1, 2, 3, 4, 5, and 7 separately as greyscale images.
- Mark the FCC where you have opened the training fields before.
- Go to Multivariate Analysis >Supervised Classification > Maximum Likelihood.
- Pressing the shift button, mark the un-stretched band number 1, 2, 3, 4, 5, and 7 of Landsat 2001.
- Set the threshold to 0 (zero) so that all pixels get classified
- Press OK
- The process will now run for several seconds to one minute, depending on the image size, the number of bands, and your computer.
| |
|  | Nyiragongo and Nyamuragira IntroductionChange detection and hazard assessment Worksheet introductionExercise 1: Exploring and learning about the volcanoExercise 2: Lava Detection using Supervised ClassificationExercise 3: Multitemporal Change Detection and MonitoringEduspace - Download Nyiragongo_Landsat.zipEduspace - Software LEOWorks 3LEOWorks 3 Tutorial
|