Intro to Change Detection

Question 1

use cases for change detection

Answer 1

1. deforestation
2. urbanization
3. natural disaster impact assessment

Question 2

control for what incidental changes to highlight real changes of interest - 4 things

Answer 2

1. spatial registration
2. atmospheric correction
3. phenological(seasonal or natural phenomena) 4. illumination conditions

Question 3

two main classes of change detection

Answer 3

binary(yes/no change)
detailed(what changed)

Question 4

simple approaches to change detection - 4 things

Answer 4

differencing, ratios, multispectral differencing, change vector analysis

Question 5

threshold selection

Answer 5

selecting a different threshold will result in a difference in the total amount of mapped change

Question 6

How is classification change detection performed

Answer 6

simply conduct a land cover/land use classification on each of the images at different dates, and then compare the two maps

Question 7

advantages of classification change detection - 2 things

Answer 7

1. powerful not only because it can provide the detailed change matrix information, but also because it doesn’t require a threshold selection
2. high quality training data appropriate for each image, we will be able to overcome some differences that may result from atmospheric, illumination, or phenological differences

Question 8

binary change map - 2 things

Answer 8

• special type of classification map whose accuracy can be assesed with the confusion matrix approach
• wealth of accuracy measures designed to access accuracy