Skip to end of metadata
Go to start of metadata

Classify

Icon

Classify, is the process of producing meaningful thematic maps from raster data; in other words; it can be defined as grouping objects forming a specific class in a data set based on their similarities and by selecting them according to their properties.

After "Classification" operation, as the classes desired to be obtained can be associated with the image properties (controlled), they can also be specified statistically (uncontrolled).
Icon

Uncontrolled (unsupervised) Classification is a method preferred when location information on the image data is not sufficient; preliminary information about the land cover is required in this type of classification.

Controlled (Supervised/Trained) Classification is a method applied under control of the analyst. Analyst creates the required statistical basis for classification using the preliminary information provided about the land cover of the study area and the classification is performed on this basis.
Icon
 

Description

Raster

Image directory information to be classified is viewed. Raster data can be changed by using '...' button.

Method

The method desired to be used in the classification process is selected from the list.

Two different methods; controlled and uncontrolled classification can be used. While 'ISODATA' approach is used for uncontrolled classification, 'Closest Distance'', ''Highest Probability' and 'Spectral Angle' are used for controlled classification.

Tolerance

Enables the pixels to be included in the class according to the proximity of selected pixels.

Tolerance values vary based on the method selected. While ISODATA and Closest Distance is 0.1-255; a value ranging from 0 to 3.14 can be defined for Spectral Angle.

Iteration

Number of iterations to be made in uncontrolled classification method (ISODATA).

Number of Classes

Number of classes required to be formed is defined when uncontrolled classification method is used. 'Add Class' in the right click operations on the right side of the window is used for the number of classes in controlled classification methods.

Criteria Color

The color of the class set created in the classification process is defined.

If uncontrolled classification method is used, criteria color of the created class is determined by the program; in classes created by using any controlled classification method, criteria color is determined by the analyst.

Color is selected from the palette or from raster using 'Criteria Colour' in the right click list of the class line created.

Display Color

The color that will be used to show the class set on final data after classification is determined.

The average color value specified for Class Criteria Color is defined automatically as the Display Color. Color can be selected from palette with 'Display Color' operation under users right click list.

Description

Class name, type, etc. information are defined. Entered information is viewed under raster data palette tab and in the slots.

Pixel/Area (m2)/%

Pixel, area and % values of the class set are viewed. Number of pixels, area and % values represent the numbers in the preview window.

Original Image

Preview of the image selected to be used in the classification process.

Classified Image

Preview of the final image obtained after the classification process. After the class definitions are made final data preview is created with the 'Classify' button.

Icon
 For classification operations (?)

 

  • No labels