inverseProjection

Reconstruct data cube from principal component bands

collapse all in page

    Syntax

    reconstructedData = inverseProjection(pcDataCube,coeff)

    Description

    Add-On Required: This feature requires the Hyperspectral Imaging Library for Image Processing Toolbox add-on.

    reconstructedData = inverseProjection(pcDataCube,coeff) reconstructs the original spectral bands in a hyperspectral data cube from the PCA (principal component analysis) or MNF (maximum noise fraction) transformed data cube and their related coefficients.

    Note

    The Hyperspectral Imaging Library for Image Processing Toolbox™ requires desktop MATLAB®, as MATLAB Online™ and MATLAB Mobile™ do not support the library.

    example

    Examples

    collapse all

    This example uses:

    Open Live Script

    Read a hyperspectral data into the workspace.

    hcube = imhypercube("indian_pines.dat");
datacube = gather(hcube);

    Extract 10 principal component bands and the transformation coefficients using the principal component analysis method.

    [pcDataCube,coeff] = hyperpca(hcube,10);

    Reconstruct the original data from 10 principal component bands.

    reconstructedData = inverseProjection(pcDataCube,coeff);

    Display the first 10 spectral bands in input data cube.

    figure
montage(datacube(:,:,1:10),BorderSize=[10 10],Size=[2 5],DisplayRange=[]);
title("Original Data Cube")

    Figure contains an axes object. The hidden axes object with title Original Data Cube contains an object of type image.

    Display the first 10 spectral bands in the reconstructed data cube.

    figure
montage(reconstructedData(:,:,1:10),BorderSize=[10 10],Size=[2 5],DisplayRange=[]);
title("Reconstructed Data Cube")

    Figure contains an axes object. The hidden axes object with title Reconstructed Data Cube contains an object of type image.

    Input Arguments

    collapse all

    PCA or MNF transformed data cube, specified as a 3-D numeric array of size M-by-N-by-P. The PCA or MNF transformed data cube of a hyperspectral data cube is computed using the hyperpca or hypermnf functions respectively. P specifies the number of principal component bands in the transformed data cube.

    Data Types: single | double | int8 | int16 | int32 | int64 | uint8 | uint16 | uint32 | uint64

    PCA or MNF coefficients, specified as a matrix of size C-by-P.

    Data Types: single | double | int8 | int16 | int32 | int64 | uint8 | uint16 | uint32 | uint64

    Output Arguments

    collapse all

    Reconstructed data cube, returned as a 3-D numeric array of size M-N-by-C. The data type of the reconstructed data cube is same as that of the transformed data cube at the input.

    Data Types: single | double | int8 | int16 | int32 | int64 | uint8 | uint16 | uint32 | uint64

    Version History

    Introduced in R2020a

    See Also

    |