499c Lab 7 (Image Analysis): Todd Mitchell

Todd Mitchell

GEOG-499c

Lab 7: Image Analysis

1. What is the difference between spatial subsetting and spectral subsetting?

Spatial subsetting allows one to isolate a spatial copy of the original image. Spectral subsetting allows one to isolate a band or band combinations from the original image.

2. Whatís georeferencing?

Georeferencing is the process of assigning spatial coordinates to corresponding cells in a raster (image-to-map) or points in a non-referenced vector (map-to-map). The projected data in a shape file can be used to align a remotely sensed image, by choosing control points in the image and mapping them to corresponding points in the shape file.

3. In exercise 7, compare the input and output image (Hand-in 7). What does the output image tell you?

This shows the maximum and minimum areas of vegetation. The white areas correspond to rich vegetation in the original image, whereas the black areas correspond to areas of little vegetation.

4. In exercise 7, compare the two output images (Hand-in 7 and Hand-in 8), what differences can you observe?

The first output was a vegetative index and the second output was a normalized difference vegetative index (NDVI). The NDVI is less contrasty than the regular vegetative index, and shows dark features (water bodies) in the middle of areas of low vegetation very well. The regular vegetative index gives a general idea of overall vegetation, but normalizing the data can be helpful especially in areas where vegetation falls off drastically.

5. In Exercise 8, can you try to give the type of landuses for the five categories in output image (Hand-in 9)?

Because this was an unsupervised classification, most classifications overlap into others. However, the following generalizations can be made.

Class 001: magenta ñ water

Class 002: ochre ñ areas of low vegetation, including brush and chaparral

Class 003: red ñ residential built up, including landscaped areas

Class 004: light blue ñ rich vegetation

Class 005: green ñ heavy built up areas (and areas with the most reflectivity from the original data)

Maps:

Hand-In 1