Page123

 
Inspect the SPOT image again.

10. Give a short explanation of how the colours of the image can be used to map the flooded areas.  
 
Supervised classification

You will now classify the SPOT image. Open a new RGB view and consider why it could be helpful to cut out some clouds in the east as well as the lake in the south using the “crop” function. Only two classes are interesting in this case: flooded and not flooded areas. You will perform a supervised classification. The first step is to select training fields for the following features:

Flooded
Not flooded
Lake Pontchartrain
 
 

Classified image

After selecting these training fields, use the “Maximum Likelihood” algorithm to classify the image, setting the threshold at 5.

 
 

	Classified image with 3x3 post classification filter applied

Now you can apply a post classification filter, either 3X3 or 5X5. You can also adjust the LUT in order to generate a more natural looking classified image, and to facilitate interpretation. You can then add a legend.

 
 

Classified image with legend

Inspect the classified image. Evaluate the quality of the classification.

Compare your classified map with maps of the flooding found on the Internet.

Page123