Subsections

The Main InfoPACK Menus

The easiest way to access the InfoPACK library is via the menus in the main InfoPACK GUI window. Clicking on menus will generally bring up a dialog box from which files are selected, parameters are chosen and the routine is run. All menu items call the appropriate InfoPACK script. All InfoPACK scripts work file-to-file, to enable the processing of arbitrarily large images within memory. As such it is always necessary to provide an output file name in the appropriate dialog box. All menu commands are executed from within the Python shell. When a command is processed you will see in the shell window the script you have called and the parameters you are using. For ease of use, all recently accessed image file names are stored which can be accessed from a ``recent files'' button in dialog boxes and the variable ``files'' in the shell window.

The main toolbar on the GUI screen includes the following menus:

File

(see section 5.3.1) with menu items:

Open Image
Import Image
Export Image
New Script
Open Script
New Console
Exit

Tools

(see section 5.3.2) with menu items:

Create New Image
Layer Operations
Extract Sub-Image
Histogram Image
Image Statistics
Information

Image

see section 5.3.3

Segment
Texture Segment
Merge
Colseg
Classify
Change Detect
Filter
Despeckle
Edge
Threshold
Image Arithmetic

Help

Gives access to the InfoPACK and Python documentation:

InfoPack: This documentation, including documentation for all InfoLIB routines. Most users will need only this manual.
Python: For details of the base Python language.
Numeric Python: The manual for the Numeric Python extensions (used by InfoPACK).
About: Gives the version number for this release.

The File Menu

This provides access to the following facilities:

`Open Image`

Open an image file in a format supported directly by the OpenEV Image Viewer (see Section 5.4).

`Import Image`

Convert an image into the InfoPACK standard .nc format (see chapter 3). A wide range of standard image formats is supported.

`Export Image`

Convert a .nc image file to an alternative format. Full conversion to the following formats is provided:

GeoTiff (*.tif)
ERDAS Imagine (*.img)
Atlantis MFF (*.hdr)

Other export formats, including postscript and .png, are made available from the OpenEV viewer: Use the Print option and print to file.

`New Script`

This opens a blank IDLE shell window. From the file menu, selecting ``New Window'' brings up an empty IDLE editor window, ready for you to type in a Python script: a text file containing a list of InfoPACK (or other Python) commands to be executed. A Python script files are generally give the suffix .py. For details of the editor features, see section 5.7.

`Open Script`

This will prompt for the script file to be opened, and then open it in a new IDLE editor window. An IDLE shell window is also opened for the script output when run from the editor window. For details of the editor features, see section 5.7.

`New Console`

This will open a new InfoPACK shell window.

`Exit`

Exit InfoPACK.

The Tools Menu

`Create New Image`

Create an image with artifical speckle generated from a specified distribution. Useful when creating test images. Full function specifications are given in Appendix B.

`Layer Operations`

This provides access to operations for generating a 3-D (layered) image from a set of single layer images, and for extracting individual layers from a 3-D image. Two options are provided:

Split Layers: Splits a 3-D image into a numbered set of the individual layers.
Combine Layers: Combine a set of individual files into a single 3-D image.

Full function specifications are given in Appendix B.

`Extract Sub-Image`

Extract a subimage from a given image. You will be prompted for the subimage size (width and height), and for the w (width) and h (height) offsets. Full function specification is given in Appendix B.

`Histogram Image`

Produce a histogram showing the distribution of amplitudes/intensities of an image. For multi-layered images, the histogram for individual layers can be viewed as well as for the whole image data. This utility can be useful when planning to carry out supervised classification of an image. Full function specification is given in Appendix B.

`Image Statistics`

Provides a set of statistics on the specified image:

Image size and datatype
Minimum and maximum pixel values
Mean and standard deviation of pixel values
The looks measure and average local looks measure
Order (of equivalent gamma distribution) and average local order

Full function specification is given in Appendix B.

`Information`

Provides information on the selected image, including:

Size
Coordinates of origin, image corners and image centre
Pixel Datatype for each layer (``Band'')

The Image Menu

`Segment`

Segment a given image file. You will be prompted for the segmentation parameters; use the default values unless you have reason to alter these. See section A.4.1 for details of the routine parameters, and section 4.1 for a longer discussion.

`Texture Segment`

Segment a given image file using texture instead of intensity information in the image. You will be prompted for the segmentation parameters; use the default values unless you have reason to alter these. See section A.4.2 for details of the routine parameters, and section 4.1.3 for a longer discussion.

`Merge`

Merge adjacent regions in a segmentation that are not significantly different in intensity at the specified level of significance. Merge requires the segmented image, the original image, and a significance level. See section A.4.3 for details of the routine parameters, and section 4.1.2 for a longer discussion.

`Colseg`

Given a segmentation image and the original imge, colour the segmentation (ie floodfill each segment) with a value computed from any of a set of measures. See section A.4.4 for details of the routine parameters, and section 4.1.4 for a longer discussion.

`Classify`

Classify an image into a given number of classes. Classify requires a segmented image, the original image and optional means and orders images for supervised classification. See section A.4.6 for details of the routine parameters, and section 4.2 for a longer discussion.

`Change Detect`

Detect changes between layers of a multiple-layer image: for example, a multi-temporal sequence. The routines accessed look for large scale changes between layers of the specified multi-layer image; you need to supply also the segmented image.

Two options are provided:

: Change Detect: produces a single layer output image denoting areas in which changes were found between any of the images. See section A.4.7 for details of the routine parameters, and section 4.3.1 for a longer discussion.
: Change Position Detect: detects changes between adjacent layers. The output is an image with one less layer than the input image; each output layer contains the changes found between one adjacent pair of layers in the input image. This option is useful for example when monitoring the growth of crops over the seasons. See section A.4.8 for details of the routine parameters, and section 4.3.2 for a longer discussion.

`Filter`

This provides access to a range of low level filters of the type traditionally used in optical image processing. See section A.4.10 for details of the routine parameters, None of the high level routines in InfoLIB use these filters; they are provided primarily as a convenience.

Single image functions: Provides access to the following arithmetic operations on images, carried out on each pixel:
- Scalar Add: image2 = image1 + c
- Scalar Multiply: image2 = image1*c
- Scalar Divide: image2 = image1/c
- log10: image2 = $\log_{10}(\mbox{image1})$
- log e: image2 = $\log_{e}(\mbox{image1})$
- square root: image2 = $\sqrt{\mbox{image1}}$
- square: image2 = image1*image1
Double Image Functions:
Provides access to the following operations, carried out on each corresponding pair of pixels in two specified images of the same size:
image3 = image1 op image2, op = +,-,*,/