14014 Rate this article:

Beyond NDVI

Rebecca Lasica

Precision Agriculture is a cutting edge topic in the remote sensing world. With several new sensors and platforms generating higher spatial and spectral resolution data, more and more growers are compelled to delve into the remote sensing realm. Generating whole-field metrics rather than extrapolating sparse ground measurements to a large area is one of the ways these growers are able to achieve better yield predictions and attain actionable information during the growing season to maximize efficiency.

One great way to generate whole field metrics is to calculate a vegetation index (Vi) to extract information about an area of interest (AOI). For example, perhaps you want to calculate the overall vegetation health within a field. Probably the most popular way to achieve this is to calculate a normalized difference vegetation index (NDVI). This index is widely accepted as a standard to determine vegetation health. Bright areas indicate healthy vegetation and darker areas are unhealthy vegetation or bare or impervious surfaces.

But NDVI is certainly not the only index out there. Many other indices lend insight into field health by utilizing spectral information in such a way that soil effects can be minimized, subtle changes within a crop can be detected, and even yield estimates can be fine-tuned. Here are just afew examples of some of the spectral indices that are useful tools for precision agriculture applications.

DVI: Difference Vegetation Index

DVI = NIR - Red

DVI is useful to differentiate between soil and vegetation. When considering a precision agriculture application and the very high spatial resolution images that are sometimes available for this analysis, it is important to mask or remove bare soil prior to running a vegetation health index so that the index does not over saturate and become difficult to interpret. The DVI does a great job separating bare soil from planted crops.

QuickBird image courtesy of Digital Globe. Data processed to reflectance then (left) calculated single field NDVI with yellow to green color ramp applied to indicate vegetation health and (right) calculated DVI and threshold to separate bare soil (blue) from vegetation (green)

GVI: Green Vegetation Index (Tasseled Cap)

GVI = (-.2848*TM1) + (-.2435 * TM2) +(-.5436*TM3) +(.7243*TM4)+(.0840*TM5)+(-.1800*TM7)

The Tasseled Cap, or Green Vegetation Index, is a great way to not only minimize the effects of background soil, but also to simultaneously emphasizing the reflective characteristics of green vegetation. While this index was originally developed for use with Landsat TM, it also works with the corresponding wavelengths found in Landsat ETM+ and Landsat 8. The numbers you see associated with each band are coefficients used to weight pixel values in those bands and thus generate a transform image.

Tasseled Cap image from Landsat TM scene calibrated to reflectance. Bright areas are more green/vegetated.

MSI: Moisture Stress Index:


The Moisture Stress Index is a narrowband index (as opposed to the broadband indices previously discussed see this for more information between the two). This narrowband index requires information in the 1599 nano meter range which is less common for many multispectral sensors used in agricultural applications. I am highlighting it here because water content measurement is important to several precision agriculture applications.In general, as water content increases, spectral absorption at 1599 nm also increases and if this information is available from the sensor, it can be useful to perform canopy stress analyses, perform productivity predictions, and even generate fire hazard condition analyses. With the introduction of UAS and the ability to interchange a payload, if this index would be useful in your application – be sure the sensor includes this wavelength!


RENDVI = P750-P705/ P750+P705

And finally I am highlighting RENDVI. You can see from the equation itself, it is very similar to NDVI. But instead of using thetraditional main absorption and reflection peaks (Red and NIR respectively) it uses information along the red edge. This is where spectral sensitivity will be found due to subtle changes in canopy foliage content, gap fraction, and senescence. This index is very useful in precision agriculture, forest monitoring, and detecting vegetation stress because it highlights subtle but detectable sensitivities long before changes are visible.

These indices (and many more!) are great ways to generate whole field metrics. And with new data sources enabling very high spatial resolution analyses, imagine the actionable information that can be achieved even on a “per plant” scale. 


A Technical Dive into Creating Accurate Digital Terrain Models


Live Webinar | July 25 Creating accurate digital terrain models (DTMs) is essential for modeling areas of... more »

The Science, and Art, of Hyperspectral Image Analysis


Harris Geospatial Solutions – June 7, 2018 Automation is being touted as the wave of the future for... more »


Sign up to receive the latest news, events, technologies and special offers.