First, verify that your two images (before [T1] and after [T2]) are correctly registered or use the ArcMap GEOREFERENCING TOOLS to register the images precisely because change detection to function properly, the before and after images must align exactly. If features do not align, results can be skewed.

FEATURE ANALYST has the ability to perform change detection on image data sets that show change over time. It is up to you to decide whether you want to extract all changes in the area of interest or want to employ object-specific change detection. Object-specific change detection allows you to extract changes in specific class types i.e., urban sprawl, change in vegetation cover or change in any other specific land use land cover class.

The first step in a change-detection extraction is creating a composite raster comprising all bands from the before [T1] and after [T2] images. Once you have your composite change-detection layer, you can draw a training set targeting either features added over time or features lost over time.

Follow the atached general workflow chart to run an extraction for changed features.

Preparation of Change-Detection Layer

The change-detection composite layer improves visualization by changing the colour of the features to define and enhance the differences between the two images.

1. Load your images into the table of contents.

2. Review alignment and compare common image features, which must overlap exactly.

3. Choose FEATURE ANALYST on the toolbar, choose RASTER TOOLS on the drop-down menu, and then choose PREPARE CHANGE DETECTION LAYER.

The PREPARE CHANGE DETECTION LAYER dialog box will open.

The activated images loaded in the table of contents display in the AVAILABLE RASTER LAYERS LIST BOX.

4. In accordance with chronology, select the first image in the Available raster layers list box and choose SET BEFORE [T1] IMAGE.

The first image will be displayed in the BEFORE IMAGE field.

5. Select the last image in the AVAILABLE RASTER LAYERS LIST BOX and choose SET AFTER [T2] IMAGE.

The last image will be displayed in the AFTER IMAGE field.

6. Choose OK.

The Save Raster As dialog box will open.

By default, FEATURE ANALYST names the file in the Name field using the last file name, the process (in this case CDL for Change Detection Layer), and incremental numbering.

7. Accept the default file name and file location.

OR

Provide a new name and location for the file and choose Save.

When the input raster images are both panchromatic images, Feature Analyst automatically adds an extra band to enable an RGB composite display for improved visualization.

The FEATURE ANALYST PROCESS box will open, displaying the progress of the composite layer.

The change-detection composite layer will appear in the table of contents and will be displayed in the workspace.

In the resulted image, reddish areas indicate features removed; blue-green areas indicate features added.

8. Examine the change-detection raster.

Thank you Sir for such detailed description. It is very helpful.