Understanding the main vegetation indices for agriculture

The Normalized Difference Vegetation Index (NDVI) is a measure commonly used in remote sensing to assess vegetation health and density. It's derived from satellite imagery and is calculated using the near-infrared (NIR) and red bands of the electromagnetic spectrum.

1. Acquire Satellite Imagery:

Access to multispectral satellite or aerial imagery is essential. This imagery should include bands in the red and near-infrared (NIR) portions of the electromagnetic spectrum. Common satellite sensors used for NDVI analysis include Landsat, Sentinel-2, and MODIS.

Basic knowledge of image processing techniques is beneficial for tasks such as image enhancement, geometric correction, and band combination. Understanding how to preprocess satellite imagery to account for factors like atmospheric interference or sensor calibration issues is important for accurate NDVI calculations.

Familiarity with the NDVI calculation formula is essential. NDVI is calculated using the formula, where NIR represents the reflectance in the near-infrared band and Red represents the reflectance in the red band.

NDVI=(NIR+Red)/(NIR?Red)

Where:

• NIR is the reflectance in the near infrared band

Near-Infrared (NIR) Band:

The near-infrared band spans wavelengths in the range of approximately 750 to 900 nanometers (nm). Unlike the red band, NIR light is mostly reflected by vegetation rather than absorbed. Healthy vegetation reflects a large proportion of NIR light due to the internal structure of plant leaves, specifically the presence of cell walls and air spaces. Consequently, areas with healthy vegetation appear bright in the NIR band.

• Red is the reflectance in the red band

Red Band:

The red band typically corresponds to wavelengths in the range of approximately 630 to 690 nanometers (nm). This band is sensitive to chlorophyll absorption in green vegetation. Healthy vegetation absorbs much of the red light for photosynthesis, so areas with abundant vegetation appear dark in the red band.

Who to compose:

To compose an image from Landsat or Sentinel with the Google Earth Engine API using specific bands such as red and near-infrared (NIR), you can follow these general steps:

1. Import the image collection: Import the Landsat or Sentinel image collection using the appropriate image collection ID and filter it based on your area of interest and time range.
2. Filter by metadata: Optionally, you can filter the image collection based on cloud cover or other metadata criteria to select the most suitable images for your analysis.
3. Select bands: Select the bands you need for NDVI calculation, which typically include the red and NIR bands.
4. Calculate NDVI: Compute the NDVI using the selected bands. NDVI is calculated as (NIR - Red) / (NIR + Red).
5. Visualize the image: Visualize the NDVI image to observe the distribution of vegetation across your area of interest.

//Some parts of this code can be found in the GEE API

// Define the region of interest
var roi = ee.FeatureCollection('put your feature here');

// Create an empty image to use for outlining the region of interest
var empty = ee.Image().byte();
var outline = empty.paint(roi, 1, 2);

// Center the map on the region of interest
Map.centerObject(roi);

/*************** Applying a cloud mask to the Landsat collection **********************/
function maskL8sr(image) {
    // Bit 0 - Fill
    // Bit 1 - Dilated Cloud
    // Bit 2 - Cirrus
    // Bit 3 - Cloud
    // Bit 4 - Cloud Shadow
    // Bit 5 - Snow
    var qaMask = image.select(['QA_PIXEL']).bitwiseAnd(parseInt('111111', 2)).eq(0); // Clear conditions
    var saturationMask = image.select("QA_RADSAT").eq(0); // Radiometric saturation QA

    // Apply scaling factors to the appropriate bands
    var opticalBands = image.select("SR_B.").multiply(0.0000275).add(-0.2);
    var thermalBands = image.select("ST_B.*").multiply(0.00341802).add(149.0);

    // Replace the original bands with the scaled ones and apply the masks
    return image
        .addBands(opticalBands, null, true)
        //.addBands(thermalBands, null, true)
        .updateMask(qaMask)
        .updateMask(saturationMask)
        .copyProperties(image, image.propertyNames()); // Copy the collection's properties
}

/**Importing Landsat 8 collection **/
var l8 = ee.ImageCollection("LANDSAT/LC08/C02/T1_L2")
    .filterDate('2013-01-01', '2022-12-31')
    .filterBounds(roi)
    .filter(ee.Filter.lt('CLOUD_COVER', 1))
    .map(maskL8sr);

// Print information about the Landsat 8 collection
print('1 - Landsat 8 Collection:', l8);
print('2 - Number of images used:', l8.size());

// Add layers
Map.addLayer(l8.median().clip(roi),{bands: ['SR_B4', 'SR_B3', 'SR_B2'], min: 0.022, max: 0.135},'L8')        

Using the images provided, the following function can be used to calculate the indices. I thought about adding other related indices, but you could focus solely on NDVI

/** Indices Calculation for Collection**/
function calculateIndices(image) {
    // Calculate NDVI (Normalized Difference Vegetation Index)
    var ndvi = image.normalizedDifference(['SR_B5', 'SR_B4']).rename('NDVI'); 

    // Calculate EVI (Enhanced Vegetation Index)
    var evi = image.expression(
        '2.5 * ((NIR - RED) / (NIR + (6 * RED) - (7.5 * BLUE) + 1))', {
            'NIR': image.select('SR_B5'),
            'RED': image.select('SR_B4'),
            'BLUE': image.select('SR_B2')
        }).rename('EVI'); 

    // Calculate SAVI (Soil Adjusted Vegetation Index)
    var savi = image.expression(
        '(1 + L) * (NIR - RED) / (NIR + RED + L)', {
            'NIR': image.select('SR_B5'),
            'RED': image.select('SR_B4'),
            'L': 0.5
        }).rename('SAVI');

    // Add calculated indices as bands to the image
    var imageWithIndices = image.addBands([ndvi, evi, savi]);

    // Clip the image to the region of interest
    var clippedImage = imageWithIndices.clipToCollection(roi);

    // Copy image properties and set the date property
    var finalImage = clippedImage.copyProperties(image, image.propertyNames())
        .set('date', image.date().format('YYYY-MM-dd'));

    return finalImage;
}        

NDVI (Normalized Difference Vegetation Index) values range from -1 to +1 because it is calculated as the normalized difference between the near-infrared (NIR) and red bands of remote sensing imagery. This calculation is based on the principle that healthy vegetation absorbs most of the visible light (including red wavelengths) for photosynthesis and reflects a large portion of near-infrared light. On the other hand, non-vegetated surfaces such as water, soil, and human-made structures have different reflectance properties.

Here's a breakdown of what different ranges of NDVI values represent:

1. Negative Values (-1 to 0):Negative NDVI values typically indicate features such as water bodies, exposed soil, or human constructions.Water absorbs both visible and near-infrared light, resulting in low reflectance in both bands and negative NDVI values. Exposed soil and human constructions also tend to have low reflectance in the NIR band, leading to negative NDVI values.
2. Values Close to Zero (Around 0):NDVI values close to zero generally represent areas with little or no vegetation cover.Bare soil, rocks, urban areas, and other non-vegetated surfaces often have NDVI values close to zero because they reflect similar amounts of visible and near-infrared light.
3. Values Close to +1 (0.3 to 1):Positive NDVI values, particularly those close to +1, indicate areas with dense, healthy vegetation cover.Healthy vegetation reflects a significant amount of near-infrared light while absorbing most of the red light for photosynthesis. This results in high NDVI values close to +1.Forests, grasslands, and other densely vegetated areas typically exhibit NDVI values close to +1, especially during the growing season when vegetation is actively photosynthesizing.

Applications of NDVI in Agriculture:

Since its creation, NDVI has become an essential tool in precision agriculture. Its ability to provide detailed information on vegetation allows for a number of practical applications. One of the main applications is monitoring plant health over time and space. Farmers can use NDVI to identify areas with water stress, nutritional deficiencies or pest attacks, allowing for precise and timely interventions to mitigate these problems. In addition, NDVI is widely used in irrigation management, helping farmers to optimize water use and ensure even distribution across fields.

NDVI mathematically contrasts the absorption of visible red light with the reflection of near-infrared light. In healthy plants, chlorophyll pigments absorb a significant portion of the visible red light for photosynthesis, while the cellular structure of the leaf surface (canopy) predominantly reflects near-infrared light.

This dynamic signifies that areas with robust photosynthetic activity, typically associated with dense vegetation, exhibit lower reflectance in the red band and higher reflectance in the near-infrared band.

The significance of NDVI in agriculture lies in its efficiency, rapidity, and practicality in estimating vegetation health and vigor. Unlike traditional monitoring methods, remote sensing techniques enable the detection of crop variability throughout their growth stages, facilitating timely interventions and informed decision-making for optimized crop management strategies.

What are the main types of vegetation indices used in agriculture?

For agricultural professionals looking to track crop development over time and easily identify changes in plant vigor, it is essential to master the use of vegetation indices.

Although often mentioned in agricultural insurance news, these indices are not always fully understood. They represent a variety of metrics based on NDVI. Using these indices as adjustments for factors such as soil brightness, atmospheric effects and others is crucial for a better understanding of crop monitoring and its limitations.

Among the various types of vegetation indices, there are three that are particularly important for intelligent risk management related to crop development: SAVI, EVI and VARI.

a) Soil Adjusted Vegetation Index (SAVI)

SAVI arose from the need to reduce the effects of soil on data capture. Its formula is similar to that of the NDVI, but includes the L factor, which varies according to soil cover. This index is widely used in areas with little vegetation or at the start of planting. Its values range from -1 to +1, indicating that the closer to L=0, the more vegetation, and the closer to L=1, the less vegetation.

b) Enhanced Vegetation Index (EVI)

The EVI is calculated in a similar way to the NDVI, but with some modifications that correct for undesirable reflections in the reflected light. These reflections can be caused by particles in the atmosphere and ground cover. EVI aims to improve data capture and interpretation.

c) Visible Atmosphere Resistance Index (VARI)

VARI, like EVI, was developed to correct for atmospheric effects. It is especially useful for detecting stressed areas in crops. Unlike previous vegetation indices, which are based on near-infrared (NIR) sensors, VARI images are generated using RGB composition. This technique aims to enhance the colors of the mapped targets, broadening the measurement possibilities, without replacing NDVI technology, but rather complementing it.