Multi-threshold Segmentation splits the domain based on pixel values. This creates image objects and classifies them based on user-created thresholds. It is also used to create unclassified image objects based on pixel values thresholds.
Please take a look at the video provided here that explains the algorithm and shows use cases:
Video - Multi-threshold Segmentation
Supported Domains
Pixel Level; Image Object Level; Current Image Object; Neighbor Image Object; Super Object; Sub Objects; Linked Objects
Level Settings
Level Name
Available only if the domain is pixel level. Select an existing level or enter a name.
Overwrite Existing Level
This parameter is only available when domain pixel level is selected. It allows you to automatically delete an existing image level above the pixel level and replace it with a new level created by the segmentation.
Image Layer
Select an image layer.
Ensure Connected Objects
Select the type of image objects that are created (for domain pixel level):
- Yes - creates connected image objects - with all pixels connected
- No - creates disconnected image objects
Merge Image Objects First
This parameter specifies the behavior if the Ensure Connected Objects parameter is set to Yes. It is only available if you select a domain that represents existing image objects.
- If Yes, all image objects with the same classification are first merged into one disconnected object and then separated again into connected image objects. By using this option you might lose some borders between image objects that existed before executing the algorithm.
- If No, only disconnected image objects are further separated into connected image objects where necessary.
Minimum Object Size
Minimum object size for connected image objects. If larger than 1, then fusion of smaller object will be performed after the segmentation. Disconnected objects will be fused to the greatest extent possible.
Thresholds
You can set multiple thresholds to classify image objects based on pixel values. An additional threshold field is added as each threshold is created. Multiple thresholds must be in ascending order:
Class 1
Pixels below or equal to threshold 1 are added to an image object. Resulting objects stay unclassified (default) or are classified as the class selected or entered. The option <no object> creates no image object for the selected range.
Threshold 1
Enter a pixel value below or equal pixels are classified as defined in the Class field. Alternatively, you can select a feature or a variable. To create a new variable, type a name for the new variable and click OK to open the Create Variable dialog box.
