ndvi

Normalized difference vegetation index

Syntax

output = ndvi(spcube)

output = ndvi(spcube,BlockSize=blocksize)

Description

output = ndvi(spcube) computes the normalized difference vegetation index (NDVI) value for each pixel in the data cube and returns an NDVI image. The NDVI image displays the vegetation cover regions of the input hyperspectral or multispectral data. The function computes the NDVI value using the red (R) band and the near-infra red (NIR) band images in the data cube. The ndvi function uses the 670 nm and 800 nm band reflectance values for the red and NIR band images respectively.

output = ndvi(spcube,BlockSize=blocksize) specifies the block size for block processing of the spectral data. (since R2021a)

The function divides the input image into distinct blocks, processes each block, and then concatenates the processed output of each block to form the output matrix. Spectral images are multi-dimensional data sets that can be too large to fit in system memory in their entirety. This can cause the system to run out of memory while running the ndvi function. If you encounter such an issue, perform block processing by using this syntax.

For example, ndvi(spcube,BlockSize=[50 50]) divides the input image into non-overlapping blocks of size 50-by-50 and then computes the NDVI values for pixels in each block.

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™ and MATLAB Mobile™ do not support the library.

Examples

Measure Vegetation Cover in Hyperspectral Data Using NDVI Image

Read hyperspectral data into the workspace.

hcube = imhypercube("indian_pines.dat");

Compute the NDVI value for each pixel in the data cube.

ndviImg = ndvi(hcube);

Estimate a contrast-stretched RGB image from the original data cube by using the colorize function.

rgbImg = colorize(hcube,Method="rgb",ContrastStretching=true);

Display the original and the NDVI image.

fig = figure(Position=[0 0 1200 600]);
axes1 = axes(Parent=fig,Position=[0 0.1 0.4 0.8]);
imshow(rgbImg,Parent=axes1)
title("RGB Image of Data Cube")
axes2 = axes(Parent=fig,Position=[0.45 0.1 0.4 0.8]);
imagesc(ndviImg,Parent=axes2)
colorbar
title("NDVI Image")

Vegetation regions typically have NDVI values greater than 0.2. NDVI values less than or equal to 0.2 indicate the absence of vegetation. Perform thresholding of NDVI image to segment the vegetation regions. Specify the threshold value.

threshold = 0.2;

Create a binary image by applying the threshold. The regions in the binary image with a value of 1 correspond to the vegetation regions in the data cube with NDVI values greater than the threshold. All other pixels have a value 0.

bw = ndviImg > threshold;

Overlay the binary image on to the RGB image and display the overlaid image.

overlayImg = imoverlay(rgbImg,bw,[0 1 0]);
figure
imagesc(overlayImg)
title("Vegetation Region Overlaid on RGB Image")

Compute the vegetation cover based on the total number of pixels in a spectral band and the number of pixels with an NDVI value greater than 0.2.

numVeg = find(bw == 1);
datacube = gather(hcube);
imgSize = size(datacube,1)*size(datacube,2);
vegetationCover = length(numVeg)/imgSize

vegetationCover = 
0.5696

Identify Vegetation and Non-Vegetation Spectra

This example shows you how to:

• Use 2-D spectral data as a hypercube for the hyperspectral functions.

• Separate vegetation and non-vegetation spectra by using ndvi function.

This example 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™ and MATLAB® Mobile™ do not support the library.

Load 2-D Spectral Data

Load 2-D spectral data containing 20 endmembers of the Indian Pines data set into the workspace.

load("indian_pines_endmembers_20.mat")

Load the wavelength values for each band of the Indian Pines data set into the workspace.

load("indian_pines_wavelength.mat")

Prepare Test Data to Use for Hyperspectral Functions

Reshape the 2-D spectral data into a 3-D volume data using the reshape function.

[numSpectra,spectralDim] = size(endmembers);
dataCube = reshape(endmembers,[numSpectra 1 spectralDim]);

Create a 3-D hypercube object, with a singleton dimension, by specifying the 3-D volume data dataCube and wavelength information wavelength to the hypercube function.

hcube = imhypercube(dataCube,wavelength);

Compute NDVI to Separate Vegetation and Non-Vegetation Spectra

Compute the NDVI value for each spectrum in the hypercube object.

ndviVal = ndvi(hcube);

Vegetation spectra typically have NDVI values greater than zero and non-vegetation spectra typically have NDVI values less than zero. Perform thresholding to separate the vegetation and non-vegetation spectra.

index = ndviVal > 0;

Plot the vegetation and non-vegetation endmembers.

subplot(2,1,1)
plot(endmembers(index,:)')
title("Vegetation endmembers")
xlabel("Bands")
ylabel("Reflectance Values")
axis tight
subplot(2,1,2)
plot(endmembers(~index,:)')
title("Non-Vegetation endmembers")
xlabel("Bands")
ylabel("Reflectance Values")
axis tight

Input Arguments

spcube — Input spectral data

hypercube object | multicube object

Input spectral data, specified as a hypercube or multicube object. If spcube is a multicube object, all its spectral bands must have the same data resolution. If all spectral bands of the multicube object do not have the same resolution, resample the bands using the resampleBands function, or select bands with uniform resolution using the selectBands function.

blocksize — Size of data blocks

2-element vector of positive integers

Size of the data blocks, specified as a 2-element vector of positive integers. The elements of the vector correspond to the number of rows and columns in each block, respectively. The size of the data blocks must be less than the size of the input image. Dividing the spectral images into smaller blocks enables you process large data sets without running out of memory.

• If the blocksize value is too small, the memory usage of the function reduces at the cost of increased execution time.

• If the blocksize value is large or equal to the input image size, the execution time reduces at the cost of increased memory usage.

Example: BlockSize=[20 20] specifies the size of each data block as 20-by-20.

Output Arguments

output — Output NDVI image

matrix

Output NDVI image, returned as a matrix of size M-by-N. M and N are spatial dimensions of the input data cube. If the data type of the input data cube is double, the output data type is also double. Otherwise, the output data type is single.

The function computes the NDVI value for each pixel as

$$NDVI=\frac{NIR-R}{NIR+R},$$

The values are in the range [-1, 1]. A value close to 1 indicates healthy and dense vegetation, 0 indicates unhealthy vegetation such as barren soil, and -1 indicates no vegetation such as water bodies.

Data Types: single | double

References

“Change Detection of Vegetation Cover by NDVI Technique on Catchment Area of the Panchet Hill Dam, India.” International Journal of Research in Geography 2, no. 3 (2016). https://doi.org/10.20431/2454-8685.0203002.

[1] Haboudane, D. “Hyperspectral Vegetation Indices and Novel Algorithms for Predicting Green LAI of Crop Canopies: Modeling and Validation in the Context of Precision Agriculture.” Remote Sensing of Environment 90, no. 3 (April 15, 2004): 337–52. https://doi.org/10.1016/j.rse.2003.12.013.

[2] “Change Detection of Vegetation Cover by NDVI Technique on Catchment Area of the Panchet Hill Dam, India.” International Journal of Research in Geography 2, no. 3 (2016). https://doi.org/10.20431/2454-8685.0203002.

Version History

Introduced in R2020a

R2025a: Process multispectral data specified as multicube object

Process multispectral data by specifying the input argument spcube as a multicube object.