ImggAggregateWavelengths v1¶

Summary¶

Aggregates images from multiple energy bands or wavelengths

Properties¶

Description¶

This algorithm aggregates images from multiple energy bands or wavelengths into one or more output wavelength bands. The algorithm applies to energy selective imaging data. It aggregates images by summing up counts, and assumes that the image pixel values represent neutron counts.

Given an input path (directory) and an output path (directory) the algorithm will combine images from the input path and write the result into the output path. Two different uses are considered: neutron radiography (single projection angle) and tomography (multiple projection angles). The image format supported is FITS (using the algorithm LoadFITS v1).

The algorithm produces its main outputs as files on disk. It also outputs two values: the number of projections (or angles) processed successfully (which will be always one for neutron radiography), and the number of wavelength bands aggregated. One and only one of the options UniformBands, IndexRanges or ToFRanges must be specified. This defines the number of output bands and how the input bands are aggregated or combined into the output bands.

For the sake of simplicity let us explain first the files that the algorithm produces when a single output band is generated. This single output band could correspond to the aggregation of all or a subset of the input bands. This is the case when the property UniformBands is set to 1, or the properties IndexRanges or ToFRanges are set to a single range (and not multiple ranges separated by commas).

For neutron radiography data the input path points to a set of image files where every file is assumed to correspond to a different wavelength or energy band. The algorithm will produce an image in the output path by combining the energy bands selected (as specified in the algorithm options, and using all the input images as default).

For neutron tomography data the input path points to a set of subdirectories, one per projection angle, and each of them containing image files. It is assumed that each of the directories corresponds to a projection angle. The algorithm will produce as many images in the output path as subdirectories are found in the input path a single image in the output path by combining the energy bands selected, separately for every input subdirectory, i.e., projection angle (selecting the bands as specified in the algorithm options, and using all the input bands as default). The images are written in a subdirectory that is named depending on the type of aggregation used. For the uniform bands (UniformBands property) the subdirectory will be named bands_uniform_idx_<start>_to_<end>, where <start> and <end> are the limits of the range(s). When the index ranges option (IndexRanges) is used the names of the output subdirectory will be named bands_index_idx_<start>_to_<end>. And when the time of flight range option (ToFRanges) is used the names will be bands_tof_idx_<start>_to_<end>. The initial prefix (bands_uniform, bands_index, bands_tof) can be mofified via the input properties OutputSubdirsPrefixUniformBands, OutputSubdirsPrefixIndexBands, OutputSubdirsPrefixToFBands, respectively.

The output images are created as described above when there is a single output band or stack of images. When multiple output ranges are specified the outputs are produced similarly but every range will be generated in a separate subdirectory created under the output path given to the algorithm. For example the algorithm would produce subdirectories named bands_by_index_idx_0_99, bands_by_index_idx_50_149, and bands_by_index_idx_99_149, or bands_uniform_idx_0_49, bands_uniform_idx_50_99.

A relevant consideration is how to use the range options to generate multiple output stacks of images. This is very important for performance reasons and concerns how to use the properties IndexRanges and ToFRanges. The comma separated list of ranges enables the user to generate multiple output stacks of images in one call to this algorithm. The same result can be obtained by calling the algorithm repeatedly with the different individual ranges, one at a time. This second alternative would be significantly slower though, and it is discouraged to use that approach in scripts. That is because the algorithm would need to read through the input images for every individual call. This can make a big different in terms of run time, given that this algorithm is meant to process large sets of images which can take of the order of tens of minutes or hours to read and/or write depending on the disk resources available.

As the algorithm tries to find images and subdirectories with images from the input path, to decide whether to process the data as a single projection (radiography) or multiple projections (tomography), there can be ambiguities. When an input path is given that contains both image files and subdirectories with image files, the algorithm processes the image files and does not try to process the subdirectories. That is, it is assumed that the input path given contains data for a single projection, and that the subdirectories do not necessarily contain images for different projections (tomography data).

Usage¶

Example - ImggAggregateWavelengthsSingleProjection

# Create an image combining all energy bands
projections, bands = ImggAggregateWavelengths(InputPath='D:\Data\RB000000\SampleA\',
                                              OutputPath='D:\Data\RB000000\SampleA_all_wavelenghts'
                                              UniformBands=1)

if 1 != projections:
   print "An error happened. Expected to process one projection but processed: {0}".format(projections)
# Print some details
print "Wrote a projection image combining {1} wavelength bands".format(bands)

Output:

Wrote a projection image combining 2000 wavelength bands

Example - ImggAggregateWavelengthsTomography

# Create a stack of images that can be used in tomographic reconstruction
angles, bands = ImggAggregateWavelengths(InputPath='D:\Data\RB000000\SampleA\',
                                         OutputPath='D:\Data\RB000000\SampleA\all_wavelenghts',
                                         UniformBands=1)

# Print some details
print "Wrote {0} projection images, each combining {1} wavelength bands".format(angles, bands)

Output:

Wrote 144 projection images, each combining 2000 wavelength bands

Categories: Algorithms | DataHandling\Imaging

Source¶

C++ source: ImggAggregateWavelengths.cpp

C++ header: ImggAggregateWavelengths.h