Main Content

hyperpca

Principal component analysis of hyperspectral data

Since R2020a

collapse all in page

    Syntax

    outputDataCube = hyperpca(inputData,numComponents)
    [outputDataCube,coeff] = hyperpca(___)
    [outputDataCube,coeff,var] = hyperpca(___)
    [___] = hyperpca(___,Name,Value)

    Description

    outputDataCube = hyperpca(inputData,numComponents) computes the specified number of principal components from the spectral bands of the hyperspectral data cube. The function returns a new data cube that contains the principal component bands. The number of spectral bands in the output data cube is equal to the number of specified principal components numComponents. To achieve spectral dimensionality reduction, the specified number of principal components must be less than the number of spectral bands in the hyperspectral data cube inputData.

    example

    [outputDataCube,coeff] = hyperpca(___) also returns the principal component coefficients estimated across the spectral dimension of the hyperspectral data cube.

    [outputDataCube,coeff,var] = hyperpca(___) returns the percentage of variance retained by the principal component bands in addition to the output arguments mention in the previous syntaxes.

    example

    [___] = hyperpca(___,Name,Value) specifies the principal component analysis (PCA) method and additional options by using the name-value pair arguments.

    Note

    This function requires the Hyperspectral Imaging Library for Image Processing Toolbox™. You can install the Hyperspectral Imaging Library for Image Processing Toolbox from Add-On Explorer. For more information about installing add-ons, see Get and Manage Add-Ons.

    The Hyperspectral Imaging Library for Image Processing Toolbox requires desktop MATLAB®, as MATLAB Online™ or 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 = hypercube('paviaU.dat');

    Compute the principal component bands of the hyperspectral data cube. Specify the number of principal components to extract as 10. By default, the function uses the singular value decomposition (SVD) method for extracting principal components.

    reducedDataCube = hyperpca(hcube,10);

    Display the first 10 spectral bands in input data cube.

    figure
montage(hcube.DataCube(:,:,1:10),'BorderSize',[10 10],'Size',[2 5],'DisplayRange',[]);

    For the purpose of visualization, rescale the principal component values to lie in the range [0, 1]. Display all the principal component bands extracted from the data cube.

    figure
rescalePC = rescale(reducedDataCube,0,1);
montage(rescalePC,'BorderSize',[10 10],'Size',[2 5]);
title('Principal Component Bands of Data Cube')

    This example uses:

    Open Live Script

    Read a hyperspectral data into the workspace. 

    hcube = hypercube('paviaU.dat');

    Perform PCA of input data cube using Eigen value decomposition. Specify the number of principal components to extract as 3. Derive the principal component (PC) bands, coefficients, and retained variance. 

    [outputDataCube,coeff,var] = hyperpca(hcube,3,'Method','Eig');

    For the purpose of visualization, rescale the principal component values to lie in the range [0, 1]. Display all the principal component bands extracted from the data cube.

    figure
rescalePC = rescale(outputDataCube,0,1);
montage(rescalePC,'BorderSize',[10 10],'Size',[1 3]);
title('Principal Component Bands of Data Cube')

    Plot the principal component coefficients and display the percentage of variance retained by each of the principal components. The summation of retained variance values imply that almost 99% of the information in input hyperspectral data is captured by the 3 principal components.

    figure
plot(hcube.Wavelength,coeff);
legend(['PC1';'PC2';'PC3'],'Location','SouthEast')
text(430,0.19,'Retained variance');
text(430,0.17,['PC1: ' num2str(var(1))])
text(430,0.15,['PC2: ' num2str(var(2))])
text(430,0.13,['PC3: ' num2str(var(3))])
xlabel('Wavelength')
ylabel('PC Coefficients')

    Input Arguments

    collapse all

    Input hyperspectral data, specified as a 3-D numeric array that represent the hyperspectral data cube of size M-by-N-by-C or hypercube object. If the input is a hypercube object, the function reads the data from the DataCube property of the object. The hyperspectral data cube must be real and non-sparse.

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

    Number of principal component bands to extract from the data cube, specified as a positive integer scalar. The value must be less than or equal to the number of spectral bands in the input data cube.

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

    Name-Value Arguments

    Specify optional pairs of arguments as Name1=Value1,...,NameN=ValueN, where Name is the argument name and Value is the corresponding value. Name-value arguments must appear after other arguments, but the order of the pairs does not matter.

    Before R2021a, use commas to separate each name and value, and enclose Name in quotes.

    Example: hyperpca(hcube,10,'Method','eig')

    Method for PCA, specified as one of these values:

    • 'svd' — To derive principal components by using the singular value decomposition method.

    • 'eig' — To derive principal components by using the eigen value decomposition method.

    Data Types: char | string

    Indicator for mean centering spectral bands, specified as one of these values:

    • true or 1 — To center each spectral bands in the input data cube by subtracting the mean of spectral bands before computing the principal component bands.

    • false or 0 — To compute principal component bands without mean centering the spectral bands in the input data cube.

    Data Types: logical

    Output Arguments

    collapse all

    PCA transformed data cube, returned as a 3-D numeric array of size M-by-N-by-numComponents. The spatial dimension of the output data cube is same as that of the input data cube. The spectral dimension of the output data cube is equal to the specified number of principal components numComponents.

    If the input data type is double, the output data type is also double. Otherwise, the output data type is single.

    Data Types: single | double

    Principal component coefficients, returned as a matrix of size C-by-numComponents. C is the number of spectral bands in the input data cube. Each column of coeff contains the coefficients for one principal component. The columns are in the order of descending component variance.

    If the input data type is double, the data type of coeff is also double. Otherwise, the data type is single.

    Data Types: single | double

    Variance retained by each principal component, returned as a vector of length equal to numComponents. The retained variance specifies the total percentage of variance explained by each principal component.

    If the input data type is double, the data type of var is also double. Otherwise, the data type is single.

    Data Types: single | double

    Version History

    Introduced in R2020a

    See Also

    | |