Image Classification I (Ch11: 319-348)

1. Unsupervised classification is said to be the identification of natural groups or structures within multispectral data; what operation is at the heart of unsupervised classification, how might a user interact with unsupervised classification?
2. Supervised classification can be defined as the process of using samples of known identity to classify pixels of unknown identity. What guidance does our book offer for selecting these samples?
3. What disadvantages are characteristic of supervised classification; how are user defined classes different from classes defined using unsupervised systems?
4. What are training fields? Explain why the size and uniformity of our training areas are so important?
5. Briefly discuss; fuzzy logic and fuzzy classifiers, how do these components fit into the overall scheme of classification.

1. What is digital image classification?
2. What are information classes? Spectral Classes?
3. What are the advantages of unsupervised classification?
4. Define Supervised classification.
5. Name 3 key characteristics of training areas.