Hyperspectral Image Processing

Import, export, process, and visualize hyperspectral data

Image Processing Toolbox™ Hyperspectral Imaging Library provides MATLAB^® functions and tools for hyperspectral image processing and visualization.

Use the functions in this library to read, write, and process hyperspectral data captured by using the hyperspectral imaging sensors in a variety of file formats. The library supports national imagery transmission format (NITF), environment for visualizing images (ENVI), tagged image file format (TIFF), and metadata text extension (MTL) file formats.

The library presents a set of algorithms for endmember extraction, abundance map estimation, radiometric and atmospheric correction, dimensionality reduction, band selection, spectral matching, and anomaly detection.

The Hyperspectral Viewer app enables you to read hyperspectral data, visualize individual band images and their histograms, create a spectrum plot for a pixel or region in a hyperspectral data cube, generate different color or false-color representations of hyperspectral images, and display metadata.

To perform hyperspectral image analysis, download the Image Processing Toolbox Hyperspectral Imaging Library from the Add-On Explorer. For more information on downloading add-ons, see Get and Manage Add-Ons.

Apps

Hyperspectral Viewer

Visualize hyperspectral data

Functions

expand all

Explore, Analyze, and Visualize

Read and Write

`hypercube`	Read hyperspectral data
`enviwrite`	Write hyperspectral data to ENVI file format
`enviinfo`	Read metadata from ENVI header file

Band Selection and Band Removal

`selectBands`	Select most informative bands
`removeBands`	Remove spectral bands from data cube

ROI Selection

`assignData`	Assign new data to hyperspectral data cube
`cropData`	Crop regions-of-interest

Color Transformation

colorize Estimate color image of hyperspectral data

Filtering and Enhancement

`denoiseNGMeet`	Denoise hyperspectral images using non-local meets global approach
`sharpencnmf`	Sharpen hyperspectral data using coupled nonnegative matrix factorization (CNMF) method

Data Correction

Radiometric Calibration

`dn2radiance`	Convert digital number to radiance
`dn2reflectance`	Convert digital number to reflectance
`radiance2Reflectance`	Convert radiance to reflectance

Atmospheric Correction

`correctOOB`	Correct out-of-band effect using sensor spectral response
`empiricalLine`	Empirical line calibration of hyperspectral data
`fastInScene`	Perform fast in-scene atmospheric correction
`flatField`	Apply flat field correction to hyperspectral data cube
`iarr`	Apply internal average relative reflectance (IARR) correction to hyperspectral data cube
`logResiduals`	Apply log residual correction to hyperspectral data cube
`rrs`	Compute remote sensing reflectance
`subtractDarkPixel`	Subtract dark pixel value from hyperspectral data cube
`sharc`	Perform atmospheric correction using satellite hypercube atmospheric rapid correction (SHARC)

Spectral Correction

`smileMetric`	Compute spectral smile metrics of hyperspectral data
`reduceSmile`	Reduce spectral smile effect in hyperspectral data cube

Dimensionality Reduction

`hyperpca`	Principal component analysis of hyperspectral data
`hypermnf`	Maximum noise fraction transform of hyperspectral data
`inverseProjection`	Reconstruct data cube from principal component bands

Spectral Unmixing

`ppi`	Extract endmember signatures using pixel purity index
`fippi`	Extract endmember signatures using fast iterative pixel purity index
`nfindr`	Extract endmember signatures using N-FINDR
`countEndmembersHFC`	Find number of endmembers
`estimateAbundanceLS`	Estimate abundance maps

Spectral Matching and Target Detection

`readEcostressSig`	Read data from ECOSTRESS spectral library
`sam`	Measure spectral similarity using spectral angle mapper
`sid`	Measure spectral similarity using spectral information divergence
`jmsam`	Measure spectral similarity using Jeffries Matusita-Spectral Angle Mapper method
`sidsam`	Measure spectral similarity using spectral information divergence-spectral angle mapper hybrid method
`ns3`	Measure normalized spectral similarity score
`spectralMatch`	Identify unknown regions or materials using spectral library
`spectralIndices`	Compute hyperspectral indices
`customSpectralIndex`	Compute hyperspectral index using custom formula
`ndvi`	Normalized difference vegetation index
`anomalyRX`	Detect anomalies using Reed-Xiaoli detector

Topics

Getting Started with Hyperspectral Image Processing
Basics of hyperspectral image processing.
Explore Hyperspectral Data in the Hyperspectral Viewer
This example shows how to explore hyperspectral data using the Hyperspectral Viewer app.
Hyperspectral Data Correction
Describes radiometric calibration and atmospheric correction.
Spectral Indices
Describes spectral indices.
Support for Singleton Dimensions
Analysis of 1D and 2D spectral data using singleton hypercube.

Featured Examples

Hyperspectral Image Analysis Using Maximum Abundance Classification

Identify different regions in a hyperspectral image by performing maximum abundance classification (MAC). An abundance map characterizes the distribution of an endmember across a hyperspectral image. Each pixel in the image is either a pure pixel or a mixed pixel. The set of abundance values obtained for each pixel represents the percentage of each endmembers present in that pixel. In this example, you will classify the pixels in a hyperspectral image by finding the maximum abundance value for each pixel and assigning it to the associated endmember class.

Open Live Script

Classify Hyperspectral Image Using Library Signatures and SAM

Classify pixels in a hyperspectral image by using the spectral angle mapper (SAM) classification algorithm. This algorithm classifies each pixel in the test data by computing the spectral match score between the spectrum of a pixel and the pure spectral signatures read from the ECOSTRESS spectral library. This example uses a data sample from the Jasper Ridge dataset as the test data. The test data contains four endmembers latent, consisting of roads, soil, water, and trees. In this example, you will:

Open Live Script

Classify Hyperspectral Images Using Deep Learning

Classify hyperspectral images using a custom spectral convolution neural network (CSCNN) for classification.

Open Live Script

Classify Hyperspectral Image Using Support Vector Machine Classifier

Preprocess a hyperspectral image and classify it using a support vector machine (SVM) classifier using the Statistics and Machine Learning Toolbox™.

Open Live Script

Endmember Material Identification Using Spectral Library

Identify the classes of endmember materials present in a hyperspectral image. The endmembers are pure spectral signature that signifies the reflectance characteristics of pixels belonging to a single surface material. The existing endmember extraction or identification algorithms extracts or identifies the pure pixels in a hyperspectral image. However, these techniques do not identify the material name or class to which the endmember spectrum belong to. In this example, you will extract the endmember signatures and then, classify or identify the class of an endmember material in the hyperspectral image by using spectral matching.

Open Live Script

Target Detection Using Spectral Signature Matching

Detect a known target in the hyperspectral image by using spectral matching method. The pure spectral signature of the known target material is used to detect and locate the target in a hyperspectral image. In this example, you will use the spectral angle mapper (SAM) spectral matching method to detect man-made roofing materials (known target) in a hyperspectral image. The pure spectral signature of the roofing material is read from the ECOSTRESS spectral library and is used as the reference spectrum for spectral matching. The spectral signatures of all the pixels in the data cube are compared with the reference spectrum and the best matching pixel spectrum is classified as belonging to the target material.

Open Live Script

Identify Vegetation Regions Using Interactive NDVI Thresholding

Identify the types of vegetations regions in a hyperspectral image through interactive thresholding of a normalized difference vegetation index (NDVI) map. The NDVI map of a hyperspectral dataset indicates the density of vegetation in various regions of the hyperspectral data. The NDVI value is computed using the near-infrared (NIR) and visible red (R) spectral band images from the hyperspectral data cube.

Open Live Script

Find Regions in Spatially Referenced Multispectral Image

Identify water and vegetation regions in a Landsat 8 multispectral image and spatially reference the image using the Mapping Toolbox™.

Open Live Script

Manually Label ROIs in Multispectral Image

Manually select regions of interest (ROIs) from a multispectral image and save them in a shapefile using the Mapping Toolbox™.

Open Live Script