Partitions a raster image into arbitrarily sized groups of cells from the input image based on intensity differences in the input raster image cells.
Typical Uses
- Medical image processing
- Computer Vision
- Object detection
- Recognition Tasks
How does it work?
The RasterSegmenter accepts features with raster geometry. All other geometry types will be rejected. Each input raster feature will be processed independently.
The RasterSegmenter executes the user selected segmentation algorithm on the input raster geometry and outputs one raster for each segment of each input raster.
Each output segment is a raster feature containing cell values for the segmented area and Nodata or zero values for the remainder depending on the options selected.
It is recommended to use some kind of feature enhancing pre-processing step before using the RasterSegmenter, such as the RasterSharpener or the RasterConvolver with a sharpening kernel.
Bands, Cell types, and Palettes
Rasters may have one band or multiple bands. The RasterSegmenter can handle up to 4 bands of the same datatype. Input rasters with an unexpected number of bands or with inconsistent types, will be rejected.
All raster cell types are accepted but are internally converted to Real64. If non-floating-point cell types are provided as input they will be converted to a floating-point cell type and a warning will be issued.
Alpha bands are preserved by removing them from the input raster prior to processing and appending them to the output raster after processing is complete. The output alpha band will be converted to 8 bit to match the bit depth of the RGB24 output bands.
Rasters may have no palette, one palette, or multiple palettes. For more information on raster structure, see Rasters (IFMERaster). The RasterSegmenter will resolve palettes on input rasters prior to processing. Input rasters with string palette values on cannot be resolved and will be rejected.
Selecting Raster Bands and Palettes
To select specific bands and/or palettes, use the RasterSelector prior to the RasterSegmenter.
Attributes
Attributes from the input features will be preserved in the output feature.
Segmentation Method
Watershed
Watershed segmentation partitions the image by considering the image as a function of its coordinates, where the intensity of the raster cells is the value.
A Gradient Magnitude filter is applied to the raster to enhance the borders of the image prior to running a Watershed Filter to produce an intensity based segmentation of the image.
The watershed algorithm does not produce a single image segmentation, but a hierarchy of segmentations. Different segmentations can be selected from this hierarchy by adjusting the maximum water level parameter.
Usage Notes
Choosing a Raster Transformer
FME has an extensive selection of transformers for working with raster data. They can be generally categorized as working with whole rasters, bands, cells or palettes, and those designed for workflow control or combining raster with vector data.
For information on raster geometry and properties, see Rasters (IFMERaster).
Raster Transformers
Working with Rasters
These transformers generally apply to whole rasters.
|
Sets the cell origin point within cells in a raster. |
|
|
Applies a convolution filter (sometimes called a kernel or lens) to raster features and outputs the results. |
|
|
Evaluates expressions on each cell in a raster or pair of rasters, including algebraic operations and conditional statements. |
|
|
Replaces the geometry of input raster features with a polygon covering either the extents of a raster or the extent of data within a raster. |
|
|
Extracts Ground Control Point (GCP) coordinate system and point values from a raster feature and exposes them as attributes. |
|
|
Sets Ground Control Points (GCPs) on a raster, pairing cell positions with known coordinates. |
|
|
Georeferences a raster by either known corner coordinates or origin, cell size, and rotation. |
|
|
Generates a grayscale shaded relief representation of terrain, based on elevation values. |
|
|
Alters the interpretation type of rasters, including all bands, and converts cell values if necessary. |
|
|
Merges multiple raster features into a single raster feature. |
|
|
Extracts the geometry properties of a raster feature and exposes them as attributes. |
|
|
Resamples rasters to multiple resolutions, based on either number of levels or dimensions of the smallest output raster. |
|
|
Transforms an image to minimize its difference with another. |
|
|
Resamples rasters, based on specified output dimensions, cell size in ground units, or percentage of original, and interpolates new cell values. |
|
|
Rotates a raster feature according to its rotation angle property, interpolating new cell values, updating all other affected raster properties, and producing an output raster feature with a rotation angle of zero. |
|
|
Enhances the features of a raster image. The RasterSharpener enhances the borders, lines, and curves while reducing noise in the flat areas of the raster image. |
|
|
Clips raster features using pixel bounds instead of ground coordinates, and optionally adds cells around the perimeter. |
|
|
Splits each input raster into a series of tiles by specifying either a tile size in cells/pixels or the number of tiles. |
|
|
Creates polygons from input raster features. One polygon is output for each contiguous area of pixels with the same value in the input raster. |
|
|
Creates a series of image tiles that can be utilized by web mapping applications such as Bing™ Maps, Google Maps™, or Web Map Tile Service. This is done by resampling rasters to various different resolutions and then splitting them into tiles. |
Working with Bands
These transformers generally apply to bands.
|
Adds a new band to a raster feature. |
|
|
Merges coincidental raster features into a single output raster feature, preserving and appending all bands. |
|
|
Alters the interpretation type of individual raster bands, converting cell values if necessary. |
|
|
Removes all unselected bands from a raster feature. |
|
|
Extracts the minimum and maximum band values, palette keys, and palette values from a raster feature, and adds them to a list attribute. |
|
|
Sets the band name of selected bands on a raster, making raster contents simpler to understand compared to band numbers. |
|
|
Removes the existing Nodata identifier from selected bands of a raster feature. Any values previously equal to the Nodata value are considered valid data. |
|
|
Sets a new Nodata value on selected bands of a raster feature. |
|
|
Specifies the required order of bands in a raster. Bands are reordered according to the input band indices. |
|
|
Extracts the band and palette properties of a raster feature and exposes them as attributes. |
|
|
Removes any selected bands from a raster feature. |
|
|
Separates bands or unique band and palette combinations, and outputs either individual raster features or a single new raster feature containing all combinations. |
|
|
Calculates statistics on raster bands and adds the results as attributes. |
Working with Cells
These transformers generally apply to individual cells.
|
Calculates the aspect (direction of slope) for each cell of a raster. Aspect is measured in degrees from 0 to 360, clockwise from north. |
|
|
Creates individual points or polygons for each cell in a raster, optionally extracting band values as z coordinates or attributes. |
|
|
Evaluates basic arithmetic , minimum, maximum or average operations on the cell values of a pair of rasters. |
|
|
Replaces a range of band values in a raster with a new single value. |
|
|
Rounds off raster cell values. |
|
|
Partitions a raster image into arbitrarily sized groups of cells from the input image based on intensity differences in the input raster image cells. |
|
|
Performs basic arithmetic operations on the cell values of a raster against a numeric value. |
|
|
Calculates the slope (maximum rate of change in z) for each cell of a raster. |
Working with Palettes
These transformers generally apply to palettes.
|
Creates a palette from an attribute, and adds this palette to all selected bands on a raster. |
|
|
Creates a string representation of an existing palette on a raster and saves it to an attribute. |
|
|
Generates a palette out of the selected band(s) of a raster. The output raster will have the selected band(s) replaced by a new band with a palette. |
|
|
Alters the interpretation type of raster palettes. |
|
|
Identifies the palette key that matches a raster band’s Nodata value, and sets a value on it. |
|
|
Removes selected palette(s) from raster features. |
|
|
Resolves the palette(s) on a raster by replacing cell values with their corresponding palette values. Palette values with multiple components, such as RGB, are broken down and the individual values assigned to multiple, newly-added bands. |
Workflow Control
These transformers generally control the flow of features in a workspace.
|
Forces accumulated raster operations to be processed, saving the state to disk and releasing resources to tune performance or assist with memory limitations. |
|
|
Reads raster features for testing purposes, including any accumulated raster operations. No additional operations are performed, and nothing is done with the features. |
|
|
Serializes the geometry of a raster feature into a Blob attribute, encoding the contents according to a choice of common binary raster formats. |
|
|
Creates a numeric raster of specified size and resolution, with default cell values. |
|
|
Decodes a binary attribute containing encoded rasters stored as Blobs, replacing the feature’s geometry with the decoded raster. |
|
|
Creates a color raster feature of specified size, resolution, and interpretation type, with default cell values. |
|
|
Selects specific bands and palettes of a raster for subsequent transformer operations. |
Vectors and Rasters
These transformers generally involve using rasters and vector data together.
|
Creates a raster representation of vector or point cloud input features, using the fme_color attribute over a solid background fill for vector features. Point clouds may be rendered using their color or intensity components. |
|
|
Creates a numeric raster representation of vector or point cloud input features, where cell values are taken from the z coordinates of the input features and overlaid on a uniform background. |
|
|
Generates a raster from input vector and raster features, with fine control over symbolization and labeling, using the Mapnik toolkit. |
|
|
Sets point cloud component values by overlaying a point cloud on a raster. The component values for each point are interpolated from band values at the point location. |
|
|
Extracts the band and palette values from a raster at the location of one or more input points and sets them as attributes on the feature. |
|
|
Produces a raster digital elevation model (DEM) by uniformly sampling a Delaunay triangulation generated from input points and breaklines. |
|
|
Rasterizes vector or point cloud features onto an existing raster. For vector features the fme_color attribute sets pixel color, and point clouds may be rendered using their color or intensity components. |
Configuration
Input Ports
Input
This transformer accepts only raster features. Palettes are not supported and will be resolved prior to segmentation.
Output Ports
Segmented
The output raster feature after segmenting the input raster feature.
<Rejected>
Non-raster features will be routed to the <Rejected> port, as well as invalid rasters.
Rejected features will have an fme_rejection_code attribute with one of the following values:
INVALID_PARAMETER
INVALID_RASTER_PALETTE
INVALID_RASTER_TOO_MANY_BANDS
INVALID_RASTER_NO_BAND
INVALID_RASTER_INCONSISTENT_BAND_SIZE
INVALID_RASTER_INCONSISTENT_BAND_INTERPRETATION
INVALID_RASTER_CELL_VALUE_OUT_OF_BOUNDS
INVALID_RASTER_ALL_BANDS_CONSTANT
INVALID_GEOMETRY_NAN
INVALID_GEOMETRY_TYPE
Rejected Feature Handling: can be set to either terminate the translation or continue running when it encounters a rejected feature. This setting is available both as a default FME option and as a workspace parameter.
Parameters
Watershed
Water Level Percentage
|
Maximum Height |
Floating-point percentage between 0.0 and 100.0. Sets the relative depth of metaphorical flooding of the raster. This parameter can be used to control the level of segmentation of the image. Usually, greater values of this parameter lead to a less segmented raster. The values of interest tend to be lower than 40. The default of 10 usually results in a fairly granular segmentation. Consider increasing if the scale level of the output is too segmented, and decreasing it otherwise. Water Level Maximum must be greater than Water Level Minimum. Required |
|
Minimum Height |
Floating-point percentage between 0.0 and 100.0. Sets the minimum height values, relative to the minimum and maximum intensity values in the image. It’s used to remove more shallow areas of the image that are of less interest. Greater values set a higher minimum, which discards more noise. Usually, useful values of this parameter are around one hundredth of the Maximum Water Level. (That is, if the Maximum Water level is set to 10, 0.1 is a good starting point for this parameter). The default is set to 0.1 which should clear some of the noise without affecting the prominent features of the image, but it depends heavily on the image. Water Level Maximum must be greater than Water Level Minimum. Required |
|
Gradient Magnitude Filter |
Controls how the Gradient Magnitude filter is applied. Only makes a difference with multi band rasters. Precise (Recommended): (Principal Components), Use the largest two singular values of the Singular Value Decomposition of the partial derivative. This is also the default value. This setting is slower but usually gives intuitively better results. Fast: (Rooted Weighted Sum of Derivatives). Much faster and is parallelized (works well on bigger rasters), but not as robust, it may result in a less precise segmentation. Try this if the other method is too slow. Required |
Editing Transformer Parameters
Using a set of menu options, transformer parameters can be assigned by referencing other elements in the workspace. More advanced functions, such as an advanced editor and an arithmetic editor, are also available in some transformers. To access a menu of these options, click 
beside the applicable parameter. For more information, see Transformer Parameter Menu Options.
Defining Values
There are several ways to define a value for use in a Transformer. The simplest is to simply type in a value or string, which can include functions of various types such as attribute references, math and string functions, and workspace parameters. There are a number of tools and shortcuts that can assist in constructing values, generally available from the drop-down context menu adjacent to the value field.
How to Set Parameter Values
Using the Text Editor
The Text Editor provides a convenient way to construct text strings (including regular expressions) from various data sources, such as attributes, parameters, and constants, where the result is used directly inside a parameter.
Using the Arithmetic Editor
The Arithmetic Editor provides a convenient way to construct math expressions from various data sources, such as attributes, parameters, and feature functions, where the result is used directly inside a parameter.
Conditional Values
Set values depending on one or more test conditions that either pass or fail.
Parameter Condition Definition Dialog
Content
Expressions and strings can include a number of functions, characters, parameters, and more.
When setting values - whether entered directly in a parameter or constructed using one of the editors - strings and expressions containing String, Math, Date/Time or FME Feature Functions will have those functions evaluated. Therefore, the names of these functions (in the form @<function_name>) should not be used as literal string values.
Content Types
| These functions manipulate and format strings. | |
|
Special Characters |
A set of control characters is available in the Text Editor. |
| Math functions are available in both editors. | |
| Date/Time Functions | Date and time functions are available in the Text Editor. |
| These operators are available in the Arithmetic Editor. | |
| These return primarily feature-specific values. | |
| FME and workspace-specific parameters may be used. | |
| Creating and Modifying User Parameters | Create your own editable parameters. |
Dialog Options - Tables
Transformers with table-style parameters have additional tools for populating and manipulating values.
Table Tools
|
Row Reordering
|
Enabled once you have clicked on a row item. Choices include:
|
|
Cut, Copy, and Paste
|
Enabled once you have clicked on a row item. Choices include:
Cut, copy, and paste may be used within a transformer, or between transformers. |
|
Filter
|
Start typing a string, and the matrix will only display rows matching those characters. Searches all columns. This only affects the display of attributes within the transformer - it does not alter which attributes are output. |
|
Import
|
Import populates the table with a set of new attributes read from a dataset. Specific application varies between transformers. |
|
Reset/Refresh
|
Generally resets the table to its initial state, and may provide additional options to remove invalid entries. Behavior varies between transformers. |
Note: Not all tools are available in all transformers.
Reference
|
Processing Behavior |
|
|
Feature Holding |
No |
| Dependencies | None |
| Aliases | |
| History |
FME Community
The FME Community is the place for demos, how-tos, articles, FAQs, and more. Get answers to your questions, learn from other users, and suggest, vote, and comment on new features.
Search for all results about the RasterSegmenter on the FME Community.
Examples may contain information licensed under the Open Government Licence – Vancouver and/or the Open Government Licence – Canada.