1

WORKSHOP1 ENVI : VISUALISATION AND

EXPLORATION OF LANDSAT8 IMAGE

BY

HEDIA CHAKROUN

2GHE
-----------------------------------------------------------
Software installation
Version Envi 4.5
Installer le logiciel Envi
Question on licence : Answer No , then click on Finish
Copy : D:\LOGICIELS\ENVI4.7\License.dat 
C:\Program Files (x86)\ITT\License

WARNING : this workshop has been adapted by H. Chakroun from the original
ENVI labs by P. Haverkamp & S. Hogan ( University of California Davis)

Getting Started with ENVI

- This exercise provides basic information about ENVI for your initial investigations of
the software. It is designed to introduce the basic key features of the software.
- The following exercises briefly introduce you to the graphical user interface and basic
capabilities of ENVI.
- When you start ENVI, the ENVI main menu appears as a menu bar. Clicking with the
left mouse button on any of the ENVI main menu topics brings up a menu of options,
which may in turn contain submenus with further options. The choices selected from
these submenus will often bring up dialog boxes that allow you to enter information
or set parameters relating to the ENVI function you have selected.

ENVI File Formats

- ENVI uses raster data formats consisting of a simple flat binary file and a small
associated ASCII (text) header file. Generalized raster data is stored as a binary stream
 2

of bytes in either Band Sequential (BSQ), Band Interleaved by Pixel (BIP), or Band
Interleaved by Line (BIL) format.
- ENVI also supports a variety of data types: byte, integer, unsigned integer, long
integer, unsigned long integer, floating-point, double-precision floating-point,
complex, double-precision complex, 64-bit integer, and unsigned 64-bit integer.
- The separate text header file provides information to ENVI about the dimensions of
the image, any embedded header that may be present, the data format, and other
pertinent information. The header file is normally created (sometimes with your
input) the first time a particular data file is read by ENVI. You can view and edit it at a
later time by selecting File → Edit ENVI Header from the ENVI main menu bar, or by
right-clicking on a file in the Available Bands List and selecting Edit Header. You
can also generate ENVI header files outside ENVI, using a text editor.

A. Visualizing one band of a multispectral

image
1. Select Start → All Programs → ENVI 4.5 → ENVI.

2. Open the multispectral Landsat 8 (OLI) data file of Beni Khalled Cap Bon
region in Tunisia.

This is a Radiance images (Luminance in French), whose units are in [W.m-2.sr-1]

- Select File → Open Image File.
- Enter Data Filenames file selection dialog appears.
- Navigate to …\L_VISIR_B1234567 from the input directory and click Open. The
Available Bands List dialog that appears on your screen will allow you to select
spectral bands for display and processing as shown below.

The Available Bands List is a special ENVI dialog containing a list of all the available image bands in
all open files, as well as any associated map information. You can use the Available Bands List to load
both color and grayscale images into a display by starting a new display or selecting the display number
from the Display #N button menu at the bottom of the dialog, clicking on the Gray Scale or RGB radio
button, then selecting the desired bands from the list by clicking on the band names.
 3

3. Explore information in “Band List” window

The bands of LANDSAT8 are as follows :

Band Wavelength Useful for mapping
Band 1 – coastal
0.43-0.45 coastal and aerosol studies
aerosol
Bathymetric mapping, distinguishing soil from
Band 2 – blue 0.45-0.51 vegetation and deciduous from coniferous
vegetation
Emphasizes peak vegetation, which is useful for
Band 3 - green 0.53-0.59
assessing plant vigor
Band 4 - red 0.64-0.67 Discriminates vegetation slopes
Band 5 - Near
085.-0.88 Emphasizes biomass content and shorelines
Infrared (NIR)
Band 6 - Short-wave Discriminates moisture content of soil and
1.57-1.65
Infrared (SWIR) 1 vegetation; penetrates thin clouds
Band 7 - Short-wave Improved moisture content of soil and
2.11-2.29
Infrared (SWIR) 2 vegetation and thin cloud penetration

4. Display B1 in gray scale : (for a clear visualization : Enhance Gaussian)

 4

B. Familiarizing Yourself with the Displays

When the image loads, an ENVI image display appears on your screen. The display group
consists of a Main Image window, a Scroll window, and a Zoom window. These three
windows are intimately linked; changes to one window are mirrored in the others:

The Scroll Window

 5

The Scroll window displays the entire image at reduced resolution (subsampled). The
subsampling factor is listed in parentheses in the window Title Bar at the top of the image.
The highlighted scroll control box (red by default) indicates the area shown at full
resolution in the Main Image window.

The Main Image Window

The Main Image window shows a portion of the image at full resolution. The zoom control
box (the colored box in the Main Image window) indicates the region that is displayed in
the Zoom window.

The Zoom Window

The Zoom window shows a portion of the image, magnified the number of times indicated
by the number shown in parentheses in the Title Bar of the window. The zoom area is
indicated by a highlighted box (the zoom control box) in the Main Image window.

1. Display the Near infra-red band

2. Color water bodies in blue

Tools Color Mapping  density slice
Explore the NG value for water
Use this value in “Edit Range” to make a unique class of water (that you color in blue)
 6

The result should look like this:

3. Displaying the Cursor Location and Value

- To access the shortcut menu in any display window, right-click in the window
- The cursor location and value can be obtained simply by passing the cursor over
the Main Image, Scroll, or Zoom windows. The Cursor Location/Value dialog
displays the location of the cursor in pixels starting from an origin in the upper-
left corner of the Main Image window.
- Display a value of radiance within the sea in B1 (443 nm)
 7

4. Using the Pixel Locator

The Pixel Locator allows exact positioning of the cursor. You can manually enter a
sample and line location to position the cursor in the center of the Zoom window. If an
image contains georeferenced data, you can optionally locate pixels using map
coordinates.
- From the Image window menu bar, select Tools → Pixel Locator
- Place the cursor in any of the three windows of the image display group and click
the left mouse button.
- Click the toggle button next to the projection field to toggle between true map
coordinates and latitude/longitude geographic coordinates. You can also choose
to change the selected projection by clicking the Change Proj button.
- From the Pixel Locator dialog menu bar, select File → Cancel to close the Pixel
Locator dialog.
 8

C. Enhancement of image displaying

1. Applying a Contrast Stretch
- Select Enhance from the Main Image display menu bar to display a list of six
default stretching options for each of the windows (Image, Zoom, Scroll) in the
display group.
- Select an item from the list (for example, Enhance → [Image] Equalization to
apply a histogram equalization contrast stretch to the Image display). This action
also updates the Scroll and Zoom windows of the display group. Try applying
several of the different available stretches.
Show example of results and interpret them

2. Interactive stretching and histogram visualization

Alternatively, you can define your contrast stretch interactively by selecting Enhance →
Interactive Stretching from the Main Image display menu bar.
By default, ENVI displays images with a 2% linear contrast stretch:

Compare the histograms and the displayed images in these 2 cases:

“Linear 0-255” and “linear 2%”.
Explore the other methods of enhancements (Gaussian, Equalization, Square root)
 9

D. Visualizing a color composite of a

multispectral image
1- Loading an RGB Image
ENVI allows you to simultaneously display multiple grayscale and/or RGB color
composite images.
1. To load a color composite (RGB) image of the delta area, click on the Available Bands
List.
Note: If you dismissed the Available Bands List during the previous exercises, you can
recall it by selecting Window → Available Bands List from the ENVI main menu bar.
2. Click on the RGB Color radio button in the Available Bands List. Red, Green, and Blue
fields appear in the middle of the dialog.
3. Select Band 4, Band 3, and Band 2 sequentially from the list of bands at the top of
the dialog by left-clicking on the band names. The band names are automatically
entered in the Red, Green, and Blue fields. This order allows to display a true color
image.
 10

4. Click on the Display # button at the bottom of the Available Bands List to open a New
Display in which to load the RGB image.
5. Click Load RGB to load the image into a Main Image window.
- Display the green, red and infra-red bands in RGB mode to obtain a colour composition
- Vegetated surfaces appear blue because of the high radiance of vegetation in B5
- If we display B5 in Green color, the vegetated areas will appear green). This is a false
color image.
 11

E. Filters analysis
1- Loading one band in grey scale mode
2- Filter  Convolution and Morphology
3- Apply the different following filters and interpret their effects on the image

F. NDVI calculation
1- Use the function: Transform  NDVI
2- Choose the Red and the Infra Red band and calculate he NDVI image (NDVI)
3- Use “Color Mapping” and “Histogram” to display vegetation in green and other
classes (water and soil) in white

The result should look like the following image:

12
 13

WORKSHOP2 ENVI: GEOREFERENING A

SCANNED MAP IN ENVI

BY

HEDIA CHAKROUN

2GHE
-----------------------------------------------------------
Overview of the Workshop
This workshop provides basic information about georeferencing maps in ENVI. It covers
step-by-step procedures for successful registration. It is designed to provide a starting
point to users trying to conduct image registration. It assumes that you are already
familiar with general image and maps registration and resampling concepts.

A. Display the map to georeference

1- Display the map “nefzane”
2- No reference system is displayed.
3- Explore the coordinate system information at the bottom of the map :

Customize map projection

All the projection systems are not included in the ENVI software, we are going to set the
projection system to the “Projection conique conforme de Lambert Nord Tunisie”.
Map Customize map projection
Give the following parameters, call the new projection “CCL Tunisia”
 14

In the same window:

Projection Add new projection
File Save projection to (overwrite the existing file):
C:\Program Files (x86)\ITT\IDL71\products\envi47\map_proj\map_proj.txt

If this does not work, then, edit the map_proj.txt file and add this line:
4, 6378206.4, 6356583.8, 36.000000, 9.900000, 500000.00, 300000.00, 37.550269,
34.416775, Carthage, Tunisia con conf

Ground Control Points Selection

1- Map Registration Select GCPs Image to Map. You must specify the
projection (Tunisa con conf) that should appear now in the list of projections and
set the resolution to 2m.
 15

2- Click OK to start the registration. This opens the Ground Control Points Selection
dialog Individual ground control points (GCPs) are added by entering the
coordinate of the first point (420000, 420000) and positioning the cursor position
in the two zoom images to the same ground location.
 16

Enter the 9 following points and position the cursor for each point to get its image
coordinates.

File Save GCP’s w map coord system : “NEFZANE_GCP.PTS”

If you quit and re-enter, you can close and restore this GCP file :
File Restore GCPs from ASCII

RMS calculation
 17

Press “Show List” button to verify the RMS error based on the GCP :
After you are done selecting the 9 pairs, click on individual GCPs in the Image to Map
GCP List dialog and examine the actual and predicted coordinates, and the RMS error.
Resize the dialog to observe the total RMS Error listed in the Ground Control Points
Selection dialog. In the GCP list you can order points by error (Options→Order Points
by Error) to see which GCPs are contributing the most to your RMSE.

Working with GCPs

The position of individual GCPs can be edited by selecting the appropriate GCP in the
Image to Map GCP List dialog and editing in the Ground Control Points Selection dialog.
Either enter a new pixel location, or move the position pixel-by-pixel using the direction
arrows in the dialog.
Clicking on the On/Off button in the Image to Map GCP List dialog removes selected
GCPs from consideration in the Warp model and RMS calculations. These GCPs are not
actually deleted, just disregarded, and can be toggled back on using the On/Off button.
Reduce RMS error to a value < 0.5 while keeping uniformly distributed GCP over the
entire map.
 18

Warp map
In the GCP window: Options  Warp File
Choose parameters of warping as follows:

Display the georeferenced map

Check the coordinates displayed by “Cursor Location Value”
 19

 Choose a second order polynomial method and check the result

Choose another resampling method and check the result

 20

WORKSHOP3 ENVI: CLASSIFICATION OF

LANDSAT8 IMAGE FOR LANDUSE MAPPING

BY

HEDIA CHAKROUN

2GHE
-----------------------------------------------------------
WARNING : this workshop has been adapted by H. Chakroun from the original
ENVI labs by P. Haverkamp & S. Hogan ( University of California Davis)

Overview of the Workshop

Classification is the process of assigning class membership to a set of samples. In the case
of remote sensing, the samples are the pixels of an image. Pixels are classified on the basis
of spectral similarity, using a variety of statistical techniques. Classes may be defined a
priori, using ground reference data and knowledge of the site, or may be specified from
the natural spectral groupings within an image. This tutorial leads you through a typical
multispectral classification procedure using Landsat8 OLI data from the Cap Bon region.
Results of unsupervised classifications are examined and post classification processing
including clump, sieve, combine classes, and accuracy assessment are discussed.

A. Open and Display Landsat8 OLI Data

1. Open the multispectral Landsat 8 (OLI) data file of Beni Khalled Cap Bon region in
Tunisia.

This is a Radiance images (Luminance in French), whose units are in [W.m-2.sr-1]

1. Select File → Open Image File.
2. Enter Data Filenames file selection dialog appears.
3. Navigate to …\L_VISIR_B1234567 from the input directory and click Open.
The Available Bands List dialog that appears on your screen will allow you to
select spectral bands for display and processing as shown below.
Display the green, red and near infra-red bands as follows:
 21

2. Review Image Colors

Use the displayed color image as a guide to classification. Even in a simple three-band
image, it‘s easy to see that there are areas that have similar spectral characteristics.
Bright red areas on the image have high infrared reflectance and low reflectance in the
red and green, which is characteristic of healthy vegetation, especially that under
cultivation. The urban areas of Menzel Bouzelfa and Beni Khalled are clearly
distinguishable.
 22

3. Examine Spectral Plots

Use ENVI‘s integrated spectral profiling capabilities to examine the spectral
characteristics of the data.
1. Choose Tools → Profiles → Z Profile (Spectrum) from the Main Image window
menu bar to begin extracting spectral profiles or right click in any display window and
choose Z Profile (Spectrum).
2. Examine the spectra for different areas (urban, vegetation, water bodies) and
show their different profiles. The Spectral Profile window will display the spectrum
for the pixel you selected. Note the relations between image color and spectral shape.
Pay attention to the location of the displayed image bands in the spectral profile,
marked by the red, green, and blue bars in the plot.
3. Select File → Cancel in the Spectral Profile dialog to dismiss it.
 23

B. Unsupervised Classification

Start ENVI‘s unsupervised classification routines from the ENVI main menu, by
choosing: Classification → Unsupervised → K-Means or IsoData.
K-Means
Unsupervised classifications use statistical techniques to group n-dimensional data into
their natural spectral classes. The K-Means unsupervised classifier uses a cluster
analysis approach which requires the analyst to select the number of clusters to be
located in the data. The classifier arbitrarily locates this number of cluster centers, then
iteratively repositions them until optimal spectral separability is achieved.

1. Choose Classification → Unsupervised → K-Means, use all of the following default

values :
a. Choose a subset corresponding to the zoom showed earlier
b. Choose 4 classes
c. Number of iterations: 1
d. Name the result : BK_KMEANS_4C_1IT
e. Load BK_KMEANS_4C in a new display
2- Comment the classified image

2. From the Main Image display menu, select Tools → Link → Link Display1 and Display2
click OK in the dialog to link the images.
3. Compare the K-Means classification result to the color composite image.
4. Try to identify the land cover associated with each class and write this down.
5. When finished, select Tools→ Link → Unlink Display to remove the link and dynamic
overlay.
 24

6. Use Tools  Color mapping  class color mapping, give appropriate names and
colors to the classes:
Water, Vegetation, Soil, Urban

7. Compute the areas and the % of each class :

Classification  Post classification  Class Statistics

The result should look like this:

Filename: D:\CLASS_KMEANS
Dims: Full Scene (614,400 points)

Class Distribution Summary

Water: 82,941 points (13.500%) (74,646,900.0000 Meters²)
Vegetation: 296,093 points (48.192%) (266,483,700.0000 Meters²)
Soil: 146,080 points (23.776%) (131,472,000.0000 Meters²)
Urban: 89,286 points (14.532%) (80,357,400.0000 Meters²)

8- Are regions R1 and R2 well classified?

8. Repeat classification K-means by setting the class number to 5 and the iteration
number to 5. Name the result : BK_KMEANS_5C_5IT
 25

9. Make an appropriate map of the BK_KMEANS_5C_5IT by giving the names of the

5 classes. Results should look as follows:

10. Comment the corrections

11. Compute the % of the 5 new classes