METHODOLOGY

1. The original images were cut, taking into account the municipal division and the areas of interest.

2. Bands 7, 4 and 2 were stacked and displayed in RGB.

3. The images were classified with an unsupervised Isodata clustering considering 30 clusters (groups of pixels with similar spectral responses).

4. Using the information contained in INEGI´s "Landuse and vegetation map" of 1973, 30 clusters were labeled with the following classification:

No. Name of the class

1. Forest

2. Open forest

3. Tropical Forest

4. Agriculture or pasture

5. Bodies of water

6. Bare soil

7. Others (thickets, cities, aquatic vegetation, clouds, shadows, etc.)

5. Interpretation of the image

In order to have a better approximation of the type of forest, the images were visually interpreted. "Closed forests" were selected.

Likewise, visual interpretation was used to redefine some of the classes, which showed a certain confusion, this was mainly the case for temperate forests and classes like tropical forests or agriculture.

6. Merging of results

The information obtained from non-supervised classification and from the interpretation of closed forests was merged. Some data were revised since certain areas previously defined as closed forest coincided with the tropical forest class, it was therefore necessary to apply certain conditions to define each class. Results are shown in figure 3.

FIGURE 3

Home page | Michoacán Case | Guerrero Case | Valle de Bravo Case | Naturale desasters casoe Hurricane Pauline

For doubts, questions or comments
Geographical Information Systems Sub-director