Main Content

Support for Singleton Dimensions

Hyperspectral Imaging Library for Image Processing Toolbox™ represents hyperspectral images as three-dimensional (3-D) arrays of the form M-by-N-by-P, where M and N are the spatial dimensions of the acquired data, and P is the number of spectral wavelengths used during acquisition. However, hyperspectral data can also be spectral reflectance curves obtained from ground spectrometers, and are arranged sequentially without a spatial dimension. This spectral response can be stored as a vector or a matrix. A vector of spectral data contains a single spectral reflectance curve, and a matrix of spectral data contains multiple spectral reflectance curves in sequence.

To process spectral data use these functions in Hyperspectral Imaging Library for Image Processing Toolbox, that also support hyperspectral images.

Explore, Analyze, and VisualizeData CorrectionDimensionality ReductionSpectral UnmixingSpectral Matching and Target Detection

hypercube

dn2radiance

dn2reflectance

radiance2Reflectance

empiricalLine

flatField

iarr

logResiduals

subtractDarkPixel

hyperpca

hypermnf

inverseProjection

ppi

fippi

nfindr

estimateAbundanceLS

removeContinuum

sam

sid

jmsam

sidsam

ns3

spectralMatch

detectTarget

anomalyRX

spectralIndices

customSpectralIndex

ndvi

To use 1-D or 2-D spectral data for the hyperspectral functions, you must reshape it into 3-D volume data:

  • Reshape 1-D spectral data of size 1-by-P into a 3-D hypercube of size 1-by-1-by-P, where P is the spectral dimension. For example,

    spectralDim = size(spectralData1D,2);
dataCube = reshape(spectralData1D,[1 1 spectralDim]);
hCube = hypercube(dataCube,wavelength);

    reshapes the 1-D spectral data spectralData1D of size 1-by-P into the 3-D volume data dataCube of size 1-by-1-by-P by using the reshape function. The hypercube function then creates a 3-D hypercube object by adding the wavelength information wavelength to the 3-D volume data dataCube.

  • Reshape 2-D spectral data of size M-by-P into a 3-D hypercube of size M-by-1-by-P or 1-by-M-by-P, where M represents the number of spectral reflectance curves and P is the spectral dimension. For example,

    [numSpectra,spectralDim] = size(spectralData2D);
dataCube = reshape(spectralData2D,[numSpectra 1 spectralDim]);
hCube = hypercube(dataCube,wavelength);

    reshapes the 2-D spectral data spectralData2D of size M-by-P into the 3-D volume data dataCube of size M-by-1-by-P by using the reshape function. The hypercube function then creates a 3-D hypercube object by adding the wavelength information wavelength to the 3-D volume data dataCube.

For more information on how to use 2-D spectral data for hyperspectral function, see Identify Vegetation and Non-Vegetation Spectra example.

See Also

| |

Related Topics