segmentation_SAGA_SRG: SAGA Seeded Region Growing segmentation algorithm

In joaofgoncalves/SegOptim: Optimization of Image Segmentation Parameters for Object-Based Image Analysis (OBIA)

SAGA Seeded Region Growing segmentation algorithm

Description

A function to access SAGA's Seeded Region Growing segmentation algorithm through CLI and optimize some of its parameters using genetic algorithms. The segmentation algorithm has two stages: 1) seed generation and, 2) seeded region growing.

Usage

segmentation_SAGA_SRG(
  x,
  rstList,
  outputSegmRst = NULL,
  Bandwidth = NULL,
  GaussianWeightingBW = NULL,
  VarFeatSpace = NULL,
  VarPosSpace = NULL,
  seedType = 0,
  method1 = 0,
  DWeighting = 3,
  normalize = 0,
  neighbour = 0,
  method2 = 0,
  thresh = 0,
  leafSize = 1024,
  SAGApath = NULL,
  verbose = TRUE
)

Arguments

x	A vector containing the parameters to be optimized by package GA: [1] BAND_WIDTH - bandwidth (cells; default: 10), [2] DW_BANDWIDTH - Gaussian and Exponential Weighting Bandwidth (default: 5) (both for seed generation in stage 1) and, [3] SIG_1 - variance in feature space (default: 1), [4] SIG_2 - variance in position space (default: 1) (both for region growing - stage 2).#' If you are using this function outside an optimization context it is better to directly define the segmentation parameters in Bandwidth, GaussianWeightingBW, VarFeatSpace and VarPosSpace.
rstList	File paths to the SAGA header files (format: .sgrd) used to perform image segmentation. File names should be provided either by a single string with each file separated by a ';' (e.g. red_band.sgrd;green_band.sgrd;nir_band.sgrd), or a vector of strings with one file name for each element (e.g. resulting from a list.files call).
outputSegmRst	A path to the output segmented image in SAGA .sdat/.srgd format (default: NULL; in this case the segmenter will use the working directory).
Bandwidth	Segmentation bandwidth.
GaussianWeightingBW	Gaussian and Exponential Weighting Bandwidth.
VarFeatSpace	Variance in feature space.
VarPosSpace	Variance in position space.
seedType	Seed type options: [0] minima of variance [1] maxima of variance (default: 0).
method1	Method 1: [0] band width smoothing [1] band width search (default: 0).
DWeighting	Weighting function: [0] no distance weighting [1] inverse distance to a power [2] exponential [3] gaussian weighting (default: 3) - note: only exponential or Gaussian will work! Otherwise it will fail to optimize DW_BANDWIDTH parameter.
normalize	Normalize inputs? [0] No [1] Yes (default: 0).
neighbour	Select the neighbourhood type: [0] 4 connected pixels (von Neumann), [1] 8 connected pixels (Moore) (default is 0).
method2	Method 2: [0] feature space and position [1] feature space (default: 0).
thresh	Similarity threshold (default: 0).
leafSize	Leaf size (for speed optimisation) (default: 1024).
SAGApath	Path to SAGA command-line interface (default: NULL; meaning that PATH env variable is used).
verbose	Output messages? (default: TRUE)

Details

Only accepts SAGA raster file format (problems encontered during segmentation when using other file formats). Only parameters in x will be optimized while the remaining will maintain their default values.

This function uses the following SAGA commands: saga_cmd imagery_segmentation 2 and saga_cmd imagery_segmentation 3. See more info at: http://www.saga-gis.org/saga_tool_doc/3.0.0/imagery_segmentation_2.html http://www.saga-gis.org/saga_tool_doc/3.0.0/imagery_segmentation_3.html http://www.saga-gis.org/saga_tool_doc/

Value

A list object containing output file paths resulting from the segmentation run. (these files will be cleaned after each GA iteration).

References

Adams, R. & Bischof, L. (1994): Seeded Region Growing. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.16, No.6, p.641-647.

Bechtel, B., Ringeler, A. & Boehner, J. (2008): Segmentation for Object Extraction of Trees using MATLAB and SAGA. In: Boehner, J., Blaschke, T., Montanarella, L. [Eds.]: SAGA - Seconds Out. Hamburger Beitraege zur Physischen Geographie und Landschaftsoekologie, 19:59-70. download

joaofgoncalves/SegOptim documentation built on Feb. 5, 2024, 11:10 p.m.