Discussion Questions (Week 5)

Metternicht and Zinck 2003. Remote sensing of soil salinity: potentials and constraints.

1. According to the article, around what range of wavelength are the differences between the percentage of reflectance in soil salinity distinguished the most?

2. What are some of the difficulties in remote sensing salt-affected soils and what are some of the ways to compensate?

3. Can vegetation indicate the presence of salinity in soil? Identify some of the salt-tolerant crops? What bands are these best view within?

4. How did the application of the fuzzy classifications improved the overall accuracy and what were the drawbacks?

5. Salt concentrated soils can be confused with what other features and in what bands are these primary situated?

6. Why is there not only mixed classes but also different severity levels of saline-alkaline ascertained for this study?

Shackelford and Davis 2003. A hierarchical fuzzy classification approach for high-resolution multispectral data over urban areas.

1. Why were two geographic data sets selected for this study?
2. Why were so many more reference pixels used for accuracy assessment than those used for training?
3. Why was the confusion for Grass and Tree so much different for one site than the other?
4. What is a “texture measure” as used in the paper?
5. In terms of accuracy assessment, which of the classes benefited the most from the entropy measure? Why?
6. Explain the basic idea of the length-width extraction algorithm.
7. The author’s don’t really use the term, but how would you define the term “fuzzy class”? Would Bare Soil be considered a “fuzzy class”?
8. What is a fuzzy membership vector?
9. How was defuzzification performed?
10. Why is the classifier considered “hierarchical”?

Lein 2003. Applying evidential reasoning methods to agricultural land cover classification.