Digital Image Classification

Digital image classification using ERDAS to identify the Land-use

1. Add your raster image and the google image of the same area to identify the land use, then Raster - Supervised- Signature Editor 

2. Then go to Drawing and then you can select either Grow, polygon, line as you prefer, and select 

area by looking at the google image, the add it to the signature editor and change colors as you preferred. 

3. Add as much as possible to get a more accurate result, the save the Signature editor. 

4. Then Go to Supervised- Supervised Classification 

5. Then as a input image add your image and as a signature editor select your saved editor, then as a Non-peramatic Rule assign as a Feature Space, keep other details default. "OK"

6. Add your classified image 

Geo-reference Satellite Images

How to Geo-reference Digital Images using ERDAS  

1. Open ERDAS - Add raster layer (Image)

2. Raster - Multispectral - Control points 

3. Select Polynomial 

4. Select Image Layer 

5. Then select reference image which you use to Geo-reference the image

6. Close the Window

7. Then select a place when you can find in both images, and add control Points (GCP)

8. Then add at-least add for GCP - Check the RMS error in the table, and make sure its under 1. 

9. Select "Display Resample Image Dialog" 

10. Save your out put and set the parameter as in the image

11. wait till you get 'Done' message 

Mosaic

The MosaicPro tool in IMAGINE Advantage and LPS creates and 

fine-tunes large mosaics that may include thousands of images. Offering dynamic update visualization, MosaicPro efficiently facilitates the seam-editing process.

Step 1,

Open ERDAS -- Raster -- Mosaic -- MosaicPro

Step 2,

Select "Display add Images Dialog" button to add images to be Mosaic

Step 3,

Select the Image you need..

Step 4,

To view the image press "Display Raster Images" button

Step 5,

Add another image as previous

Step 6,

Then go to Process -- Run Mosaic 

Step 7,

Select file type as you need the output 

Step 8,

Step 9,

If you need you can use this outputs in ArcGIS also..

Layer Stack

Layer Stack is use to create image with combination of Red, Green, Blue and Near IR  images. 

Step 1,

Open ERDAS -- Spectral -- Layer Stack 

Step 2,

Select Images

Step 3,

then "Add" all four images one by one in order of Red, Green, Blue, Near IR

Step 4, 

Then press "ok" to start the process..

Step 5,

Then add your Image, then go to Raster -- Multispectral -- and change color combination in to red (3) - green (2) - blue (1)

red: 0.61 - 0.69 µm

green: 0.51 - 0.60 µm

blue: 0.45 - 0.51 µm

This band combination is used to represent an image in natural colour and therefore best approaches the appearance of the landscape in reality.

Step 6,

change color combination in to (Red) (4) - red (3) - green (2)

VNIR: 0.76 - 0.90 µm

red: 0.61 - 0.69 µm

green: 0.51 - 0.60 µm 

These three bands are typically combined to make a 'traditional' false colour composite as one also knows from aerial photography. 

Data Fusion

Data fusion techniques combine data from multiple sensors, and related information from associated databases, to achieve improved ac-curacies and more specific inferences than could be achieved by using a single sensor alone.

Single sensor – spatial (QuickbirdMultispectral + Pan)
Multi sensor – spatial (Landsat 30m + SPOT 10m)

Most of these sensors operate in two modes: multispectral mode and the panchromatic             mode.
Usually the multispectral mode has a better spectral resolution than the panchromatic                   mode.
Most of the satellite sensors are such that the panchromatic mode has a better spatial                 resolution than the multispectral mode.
To combine the advantages of spatial and spectral resolutions of two different sensors,                image fusion techniques are applied  


This improves the interpretability of the data by having high resolution information which is also in color. Resolution Merge offers three techniques: Multiplicative, Principal Components, Brovey Transform and IHS under sensor merge. 

• Multiplicative is based on simple arithmetic integration of the two raster sets. 
• The Principal Components merge operates on PC-1 rather than the input raster image. 
• In the BroveyTransform option, all bands are used according to the following formula: 
• [DNB1 / DNB1 + DNB2 + DNBn] x [DNhigh res. image] = DNB1_new
• [DNB2 / DNB1 + DNB2 + DNBn] x [DNhigh res. image] = DNB2_new
• etc. 
• whereB = band 


Step 1,

Open ERDAS Imagine --- Raster --- Pan Sharpen --- Resolution Merge

Step 2,

For the High Resolution Input file select Panchromatic image.

Step 3,

For the Multispectural Input file select Multispectural image.

Step 4,

You can give output file name and in "Method" select 1st option "Principal Components".

Step 5,

Then press ok, progress window will be appear, wait few seconds to complete the process.

Step 5,

Add your image to the ERDAS

Step 6,

Then select another 2D view for next image,

Step 7,

to get Multiplicative image fallow the same steps and select "Multiplicative" option.

Step 8,

Wait for process to be complete...

Step 9,

Then add your image, the to get a Brovey Transform image followe the same steps and select "Brovey Transform"

Step 10,

Add your image,

Step 11,

Then go to Raster -- Utilities -- Sensor Merge -- 

Step 12,

Follow the same steps and select "IHS"

Step 13,

Add IHS image also to the ERDAS