Support for Singleton Dimensions

Hyperspectral Imaging Library for Image Processing Toolbox™ represents hyperspectral and multispectral images as two-dimensional (2-D) images of the form M-by-N,where M and N are the spatial dimensions of the acquired data, in P number of spectral bands. However, hyperspectral and multispectral 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 and multispectral images.

Explore, Analyze, and Visualize	Data Correction	Dimensionality Reduction	Spectral Unmixing	Spectral Matching and Target Detection
imhypercube geohypercube immulticube geomulticube	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 and multispectral functions, you must reshape it into 3-D volume data. These methods assume that, if the spectral data is multispectral data, the spatial resolution of the data is uniform across spectral bands.

• Reshape 1-D spectral data of size 1-by-P into a 3-D data cube 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 = imhypercube(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 imhypercube 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 data cube 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]);
  mcube = immulticube(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 immulticube function then creates a 3-D multicube 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.