Remote Sensing#14 - Change Detection and Error Assessment

How to work out if an area has evolved over time looking at 2 or more images. &. How to determine the level of error between those images.

In the following video for ENVI we will look at Image classification

The simplest method of change detection:
(1) The easiest way to do this is by performing all of the previous corrections and then classification of two images. One before and One after.
(2) Then create a difference image from these two (Diff = later-earlier).
(3) Divide the new difference image into classes (e.g. do another classification)
See chapter 6 of textbook - Learning Geospatial Analysis with python for reference.
Option 2:
(1) Create an error matrix table and compare image 2 to image 1 to create producer and user accuracies by putting points accross both images in either a systematic/random or stratified approach to find the percentage change of the classes those points are on in both images.