Lec 11 - Accuracy Assessment

Question 1

In the context of image interpretation, this
determines the quality of information derived from remotely sensed data.

Answer 1

Accuracy Assessment

Question 2

In this type of assessment, you determine if a map ‘looks right’ by comparing what you see in the map or image with what you see on the ground.

Answer 2

Qualitative Assessment

Question 3

Goals of Accuracy Assessment (2)

Answer 3

(1) Assess how well a classification worked.
(2) Understand how to interpret the usefulness of someone else’s classification.

Question 4

Process of Accuracy Assessment (3)

Answer 4

(1) Collect reference data: “ground truth”
(2) Determine the class types at specific locations.
(3) Compare reference to classified map.

Question 5

Sources of Data (3)

Answer 5

(1) Aerial Photo Interpretation
(2) Ground Truth with GPS
(3) GIS Layers

Question 6

Issues in using Reference Source (3)

Answer 6

(1) Extracting Relevant Information
(2) Size and Spatial Frequency of Reference Plots
(3) Positioning and Number of Samples

Question 7

Choosing Reference Source

Answer 7

Make sure to actually extract from the reference source the information that you need for the classification scheme.

Question 8

Determining Size of Reference Plots (4)

Answer 8

(1) The area of each sample site should be governed by the pixel size of the sensor and the geometric accuracy of the satellite image.
(2) Match spatial scale of reference plots and remotely-sensed data.
(3) Take into account spatial frequencies of image
(4) Check the accuracy of position of image and reference data

Question 9

Determining Position and Number of Samples (3)

Answer 9

(1) Make sure to adequately sample the landscape: represent all classes and the different environmental conditions for each class.
(2) Variety of sampling schemes:
(3) The more reference plots, the better;

Question 10

Sampling Methods (5)

Answer 10

(1) Simple Random
(2) Stratified Random
(3) Systematic
(4) Systematic Non-Aligned
(5) Cluster

Question 11

Observations are randomly placed

Answer 11

Simple Random Sampling

Question 12

Minimum number of observations are randomly placed in each category

Answer 12

Stratified Random Sampling

Question 13

Observations are placed at equal intervals according to a strategy

Answer 13

Systematic Sampling

Question 14

A grid provides even distribution of randomly placed observations.

Answer 14

Systematic Non-Aligned Sampling

Question 15

Randomly placed “centroids” used as a base of several nearby observations. The nearby observations can be randomly selected, systematically selected, etc…

Answer 15

Cluster Sampling

Question 16

Collect large sample of randomly distributed points too expensive and difficult. Only sample those pixels without influence of potential registration
error.

Answer 16

Random Sampling

Question 17

Individual pixels or clusters of pixels or polygons

Answer 17

Sample Unit

Question 18

Sample Number (3)

Answer 18

(1) General Area
(2) Large Area (or more than 12 categories)
(3) Depends on the variability of each category

Question 19

50 samples per category (scattered across the image)

Answer 19

General Area

Question 20

75-100 samples per category

Answer 20

Large area or more than 12 categories

Question 21

This needs more samples than open water.

Answer 21

Wetland

Question 22

Total Accuracy

Answer 22

Number of correct plots / Total number of plots

Question 23

Problem with Total Accuracy

Answer 23

Summary value is an average. It does not reveal if error was evenly distributed between classes or if some classes were really bad and some really good

Question 24

Error of Commission (2)

Answer 24

(1) User’s Accuracy
(2) Producer’s Accuracy

Question 25

How accurate the map is rom the perspective of the user of the classified map.

Answer 25

User’s Accuracy

Question 26

How accurate the map is rom the perspective of the maker of the classified map.

Answer 26

Producer’s Accuracy

Question 27

A measure of the difference between the actual agreement between reference data and the results of classification, and the chance agreement between the reference data and a random classifier

Answer 27

Kappa Statistic

Question 28

Accuracy Assessment Reporting (3)

Answer 28

(1) Audience
(2) Objective
(3) References

Question 29

Factors Influencing Low Accuracy (2)

Answer 29

(1) Errors in Classified Map
(2) Errors in Reference Data

Question 30

Strategies for Improving Classification (4)

Answer 30

(1) Land use/Land Cover
(2) Spectral Inseparability
(3) Atmospheric effects
(4) Scale