Section 7

Question 1

What are the four steps of a classification?

Answer 1

Select training areas
Build spectral signatures
Classification
Accuracy assessment

Question 2

Name the two approaches to digital classification.

Answer 2

Supervised

Unsupervised

Question 3

What is a supervised classification?

Answer 3

You determine the class boundary based on training sites

Question 4

What is an unsupervised classification?

Answer 4

The class boundaries are determined automatically

Question 5

What is the main source of error of classification?

Answer 5

Miss class identification by user

Question 6

What is a training site?

Answer 6

The sample areas a user selects for the classes to be based off of.

Question 7

How do you decide to use supervised or unsupervised?

Answer 7

If there is previous knowledge of the site use supervised. If not unsupervised

Question 8

In an unsupervised classification what are the two inputs by the user?

Answer 8

The preferred cluster algorithm

Number of classes

Question 9

Name the three parametric classification algorithms.

Answer 9

Minimum distance to mean

Parallelpiped

Maximum likelihood

Question 10

Name the three non parametric classification algorithms.

Answer 10

Spectral mixture analysis

Spectral angle mapper

Support vector machine

Question 11

How does the minimum distance to mean algorithm work?

Answer 11

Relies on the straight line (Euclidean) distance from the class means to the unclassified pixel. It is simple and efficient

Question 12

What is the downside to the minimum distance to mean?

Answer 12

Insensitive to different degrees of variation in spectral response of data

Question 13

How does the parallelepiped algorithm work?

Answer 13

It is the simplest method, known as the box classifier because it uses one to classify minimum and maximum ranges

Question 14

What are the problems with the parallelepiped?

Answer 14

When boxes overlap the box the is over top will claim the entire area. Like coding

Question 15

What is a stepped parallelepiped?

Answer 15

It is an improvement in the boxes are stepped to reduce overlapping and make a tighter classification

Question 16

What is the maximum likelihood classifier?

Answer 16

It uses normal distribution to make a statistical probability of the pixels location. It assigns pixels a class based on the max probability

Question 17

What is a negative of the maximum likelihood classifier?

Answer 17

The assumption that data is normally distributed and it is computationally intense

Question 18

Describe the spectral mixture analysis (SMA) algorithm.

Answer 18

It decomposes a pixel to look at the spectral signature. It can be linear or non linear

Question 19

Describe the SMA linear.

Answer 19

All reflective elements contribute to its representation. It is additive

Question 20

Describe the SMA non linear.

Answer 20

Looks at reflective components proportionally

Question 21

Describe the spectral angle mapper (SAM) algorithm.

Answer 21

It compares spectral signatures, it is insensitive to intensity

Question 22

Describe the support vector machine (SVM) algorithm.

Answer 22

Similar to maximum likelihood, it yield high classification results

Question 23

What is an error confusion matrix?

Answer 23

An accuracy assessment that does a class by class basis for relationships.

Question 24

Where is the optimal place reference data should be collected?

Answer 24

The field

Question 25

What are the five sampling strategies?

Answer 25

Systematic
Stratified
Random
Stratified random
Cluster resampling

Question 26

What is the systematic sampling strategy?

Answer 26

Uniform sampling that uses a pattern like a grid. This is not very useful.

Question 27

What is stratified sampling strategies?

Answer 27

Sample based on predetermined strategy based on location or class

Question 28

Describe random sample strategy.

Answer 28

Yields statistically selected sample with no bias

Question 29

Describe the stratified random sample strategy.

Answer 29

Most effective

Question 30

What are the downsides of cluster sampling?

Answer 30

It is only used when limited access to site and there will be bias spatial results

Question 31

What is a kappa coefficient?

Answer 31

A normalization process used to standardize error matrix results so they can be compared

Question 32

What is the key goal of a classification?

Answer 32

Simplification or reduction in the complexity of a system into something meaningful for the observer