 |
|
|
Change detection involves the use of multitemporal data sets
to discriminate change in land-cover or in a phenomenon between
dates of imaging. It is particularly useful in
Ecosystem Analysis.
Many techniques are currently being utilized to detect change
using remotely sensed data. These include:
|
Second image date highlights changes in the
jungle. (South America) |
| |
| · Image difference |
|
| · Post-classification (i.e., classify
images independently and subtract) |
| · Principal Component Analysis |
|
| · Tasseled Cap Analysis |
|
| · Vegetation Indices |
|
| · Image Ratio |
|
| · Multi-date Composite Image Change
Detection (date1 + date2 = 1 image) |
| · Manual On-Screen Digitization
of Change, etc... |
| |
|
|
Change detection between two dates requires at
least the following:
|
| · Temporal consideration |
|
| · Atmospheric correction |
|
| · Radiometric normalization |
|
| · Image registration |
|
| |
|
|
| AAI's DeltaCue
can identify small occurrences of change, over broad geographic
areas, in multispectral imagery |
| |
Examples of change that can be detected between
two dates of imagery:
|
| · Clearcutting, selective cutting,
other forestry operations |
| · New construction, development
or land scarring |
| · Spoil piles from underground digging |
|
| · Tunnel entrances and shafts |
|
| · New or expanded roads and lines
of communication |
| · Stressed vegetation |
|
| · Changes in water levels |
|
| |
|
|
