GEOG457 / 657: Lab 8: Image segmentation: Idrisi (and ASTER ?)
0. Introduction
Classification in Image processing has been conducted using per-pixel classifiers since the first Landsat in 1972. The introduction of segmentation classifiers where images are first segmented into areas of homogeneity has been described as the single biggest advance in image processing in the last 30 years. The main software has been a package known as eCognition, marketed but not produced by PCI and sold separately. software is highly regarded, - it has been described as quite complex and we have not renewed the license as it is not inexpensive. Some UNBC students have used this for projects, coincidentally only the fourth was not a CPSC major..
PCI have recently introduced their own object based classifier, which incorporates some similar featured, as in lab 6:
Idrisi ( is a popular raster GIS that has included a full suite of image processing options since its origins in 1986. The latest version named Taiga includes segmentation modules with a focus paper on the topic at:
Have a skim through this: as you can read, the process involves the use of 3 modules:
Segmentation, SEGTRAIN and SEGCLASS
This lab will take examine these steps. Note that further exploration of this process or other IDRISIimage processing options could constitute your course project.
We’ll use (again) the sub-window you created from the SPOT data in lab 1 and later labs
1. preparing the data for Idrisi
IDRISI cannot directly read a .PIX file, so the easiest format will likely be a .TIF
In PCI, use utilities -> translate to create a .TIF either keep all 4 bands, or just 2,3 and 4
You might also consider reducing its size – max 1000 x 1000
Tools -> Clipping/subsetting can reduce and save as tif in the same step
2. IDRISI TAIGA START
Idrisi is windows based software
Start Idrisi -Taiga from Start -> all programs -> IDRISI
Warning: I was raised on Idrisi versions 1-7 but lost touch with later versions (now 16), so in some areas, you may be boldly going where no UNBC’ers have boldly gone before
Briefly flip through the menu dropdowns to tally the many options
Have a brief look at help -> Quick start
Follow the instructions under Idrisi Explorer to set your project environment to a folder under your own workspace, the default otherwise is on the server where the software is located, and none of us even (especially) me can write data here
TIF file conversion / import
In the same menu, you can import your tif file for the lab: (Do it!)
Help-> ESRI quick start -> TIF import
(this option is also available from the pulldown menu through reformat-> convert)
Idrisi Raster files
The native format for raster files is .RST
Unlike PIX files, these are single bands but multiple bands can be grouped as a .RGF
To display a raster band :
Display -> display launcher
Take the defaults to display one of your bands -> OK = to the PCI pseudocolour display
Repeat, selecting the grayscale option
To display a 4-3-2 colour composite, select display -> COMPOSIT and proceed
(the band numbers may have changed during the import / convert process)
The modules we will need are under the Image Processing dropdown ->
3. SEGMENTATION
Open the help for segmentation and peruse
Select input files and output prefix, otherwise accept defaults
I had some issues here getting the inputting format right (with an error message saying it could not find band1.rst) .. in the end I’m not sure which version worked, but hope we can figure this out. When it runs, it automatically displays the vector segment it produces.
4. SEGTRAIN
SEGTRAIN creates training and signature files from a segmentation file (.VCT) created with SEGMENTATION. The user interactively creates signatures (.SCF) and assigns class names and IDs directly from the segmentation file.
SEGTRAIN uses the vector layer created in the previous step.
Follow the SEGTRAIN operations instructions in the help menu – use the same classes you may have saved in an earlier lab due to the next step (step 5).
The final result includes training signature files that will be used in classification
5. SEGCLASS
The final stage SEGCLASS is a majority rule classifier based on the majority class within a segment. It requires an already classified image and segmentation image. The classified image is derived in IDRISI using a pixel-based classifier such as MAXLIKE with the segment-based training and signature files. From what I read, MINDIST might be easier.
Follow the operations in help to run MINDIST if required … but wait …
Don’t we (you) already have a classified layer in PCI ?
If so, simply repeat the earlier steps to translate this layer to a .TIF and convert to .RST
Follow the help for SEGCLASS inputting the classification and segmentation images
------
If you are still having fun and looking for more, a sample ASTER image dataset is at: /home/labs/geog457/hoodoo.pix (named after the glacierised volcanic peak in the lower centre). Bands 1-2 are green, red, Band 3 is NIR and bands 4-9 are all MIR. View the data and if you are really gung-ho, you could segment 2-3-4, or leave it to a later project.