SAGA 9.4.2 | Tool Library Documentation

Seeded Region Growing


Description

The tool allows one to apply a seeded region growing algorithm to a stack of input features and thus to segmentize the data for object extraction. The required seed points can be created with the 'Seed Generation' tool, for example. The derived segments can be used, for example, for object based classification.


References


Parameters

 NameTypeIdentifierDescriptionConstraints
InputSeedsgrid, inputSEEDS--
Featuresgrid list, inputFEATURES--
OutputSegmentsgrid, outputSEGMENTS--
Similaritygrid, outputSIMILARITY--
Seedstable, outputTABLE--
OptionsGrid Systemgrid systemPARAMETERS_GRID_SYSTEM--
Normalize FeaturesbooleanNORMALIZEStandardize the input features, i.e. rescale the input data (features) such that the mean equals 0 and the standard deviation equals 1. This is helpful when the input features have different scales, units or outliers.Default: 0
NeighbourhoodchoiceNEIGHBOUR-Available Choices: [0] 4 (von Neumann) [1] 8 (Moore) Default: 0
MethodchoiceMETHOD-Available Choices: [0] feature space and position [1] feature space Default: 0
Variance in Feature Spacefloating point numberSIG_1-Minimum: 0.000000 Default: 1.000000
Variance in Position Spacefloating point numberSIG_2-Minimum: 0.000000 Default: 1.000000
Similarity Thresholdfloating point numberTHRESHOLD-Minimum: 0.000000 Default: 0.000000
RefreshbooleanREFRESH-Default: 0
Leaf Size (for Speed Optimisation)integer numberLEAFSIZE-Minimum: 2 Default: 256

Command Line


Usage: saga_cmd imagery_segmentation 3 [-SEEDS ] [-FEATURES ] [-SEGMENTS ] [-SIMILARITY ] [-TABLE ] [-NORMALIZE ] [-NEIGHBOUR ] [-METHOD ] [-SIG_1 ] [-SIG_2 ] [-THRESHOLD ] [-REFRESH ] [-LEAFSIZE ]
  -SEEDS:       	Seeds
	grid, input
  -FEATURES:    	Features
	grid list, input
  -SEGMENTS:    	Segments
	grid, output
  -SIMILARITY:  	Similarity
	grid, output
  -TABLE:       	Seeds
	table, output
  -NORMALIZE:   	Normalize Features
	boolean
	Default: 0
  -NEIGHBOUR:   	Neighbourhood
	choice
	Available Choices:
	[0] 4 (von Neumann)
	[1] 8 (Moore)
	Default: 0
  -METHOD:      	Method
	choice
	Available Choices:
	[0] feature space and position
	[1] feature space
	Default: 0
  -SIG_1:    	Variance in Feature Space
	floating point number
	Minimum: 0.000000
	Default: 1.000000
  -SIG_2:    	Variance in Position Space
	floating point number
	Minimum: 0.000000
	Default: 1.000000
  -THRESHOLD:	Similarity Threshold
	floating point number
	Minimum: 0.000000
	Default: 0.000000
  -REFRESH:     	Refresh
	boolean
	Default: 0
  -LEAFSIZE:    	Leaf Size (for Speed Optimisation)
	integer number
	Minimum: 2
	Default: 256