Vegetation indices in precision agriculture

A few weeks ago we started a new section on precision agriculture and we gave some hints of what this type of technology can help us in modern agriculture. Today we are going to talk about a very small part of this technological world applied to the agricultural world: the vegetation indices .

First of all, we must talk about where this term comes from and when it began to be calculated in bulk. Normally, when talking about these indices, we usually refer to the world of remote sensing. From the primitive earth observation satellites back in the 70s with the launch of the first LANDSAT to the most modern and accurate ones like the recently launched worldview-4 and not forgetting the famous drones, the world of remote sensing is on the rise.

Remote sensing is the term translated from English “remote sensing”. It is the technique that consists of capturing images of the earth at a distance in different bands of the electromagnetic spectrum and then proceeding to their processing, analysis and interpretation.

[badge style = "green"] Remote sensing is the term translated from English "remote sensing". It is the technique that consists of capturing images of the earth from a distance in different bands of the electromagnetic spectrum and then proceeding to their processing, analysis and interpretation. [/ Badge]

Satellites? Drones? Bands? Electromagnetic spectrum? But weren’t we talking about agriculture?

Patience. We will get to that. For now, some small bases have to be laid.

A LITTLE THEORY OF ELECTROMAGNETISM

We do not intend to give a first-year physics class but we do have a minimal update on what the electromagnetic spectrum consists of and how remote sensing and especially vegetation indices are based on all this.

The electromagnetic spectrum is nothing more than a division or stratification of ranges of wavelengths with similar behaviors. Recall that an electromagnetic wave is defined by:

  • Its amplitude : distance between the farthest point of a wave and the equilibrium or middle point.
  • Its wavelength : distance between two successive peaks of a wave.
  • Its frequency : number of cycles passing through the same point in a unit of time.

As we can see in the previous diagram, the different ranges in radio waves, microwaves, infrared, visible (the only portion of the spectrum that the human eye can capture), ultraviolet …

Of all this salad of electromagnetic waves, for the vegetation indices we are interested from the thermal infrared to the visible spectrum, mainly for their calculation since the plant covers, in general, respond in very specific ways in these ranges that are not useful to know its status.

WHAT IS A VEGETATION INDEX?

A vegetation index is usually the result of a formula that uses one or more bands of the electromagnetic spectrum . The relationship between these bands is behind an empirical study that shows the direct relationship between the numerical value captured by the sensor and the plant variable to be measured (normally biomass or plant vigor).

WHAT DOES IT CONSIST OF AND HOW ARE THEY CALCULATED?

This is where the whole gist of the matter is. Normally, the cameras used to calculate these indices, whether they are on board satellites, airplanes or drones, are somewhat special cameras called multispectral . Their name indicates it, they capture various spectra. For vegetation indices, they typically capture the spectrum of green, blue, red (what we call RGB), and near infrared. The blue band in these cases is not widely used and is usually replaced by the so-called near red.

Multispectral camera for drones

When the camera captures a photo in these bands, later they must be entered in a computer program and play with the formulas of the different indices to obtain values ​​that tell us something and be able to act accordingly on our crops.

HOW MANY VEGETATION INDICES ARE THERE?

A few, and constantly evolving. Companies, universities and scientists around the world are continually studying new vegetation indices, modifying existing ones to make them more accurate. There is a whole world to do, develop and discover. For now we are going to mention some that are currently being used for precision agriculture, starting with the famous and well-known NDVI.

NDVI

They are the acronym for Normalized Difference Vegetation Index. Of the vegetation indices known so far, it is one of the most important and used. It uses the red and near infrared bands and allows us to calculate the vigor of a plant in a specific phenological state. This vigor indicates the state of the crop. If it is below normal, the index will tell us that the plant is suffering  (or not at its optimum) and we will have to find out why. It may be due to water stress, a plague or disease, a lack of nutrients … The causes can be many but what the index tells us prematurely is that something is happening to the plant so that we can arrive on time and correct it.

[alert style = »green»] Do you want to know what NDVI consists of, all its fundamentals and how to apply it? Here you have an extensive article about this index. [/ Alert]

NDVI color map showing growing areas with more and less vigor

GNDVI

This is a variation of NDVI that uses the band of green instead of red for its calculations. There are crops that respond better to some bands than to others and in rice or corn crops, sometimes this index gives better results, for example, in predicting crop yield based on vigor calculated by this method.

SAVI AND MSAVI

They are the acronym for Soil Adjusted Vegetation Index. When it comes to calculating vegetative vigor, the type and state of cultivation must be taken into account. If the vegetation cover is very poor (wide planting frames or early phases of the crop) the soil reflectance indices affect the final value and this “soil effect” decompensates the NDVI or GNDVI values. For them, the SAVI or MSAVI indices (the second is a modification of the other) manage to compensate for the ground effect in these cases.

EVI Y EVI2

They are the acronym for Enhanced Vegetation Index. These vegetation indices (EVI2 is an evolution of EVI) were studied for satellite remote sensing and also use the blue band (in addition to red and near infrared) to compensate for errors derived from atmospheric distortions . Remember that the energy captured by the multispectral sensor of a satellite has to travel through the atmosphere for about 800km, which is where these types of satellites orbit. The EVI2 is a modification to avoid having to make calculations with the blue band, so it could work with camera sensors embedded in drones (which may lack the blue band).

CWSI

From the vegetation indices seen so far, this one differs a bit from the rest. This index is calculated not with a multispectral image but with a thermographic sensor that gives us the surface temperature of the culture. There are studies that demonstrate the direct relationship in the temperature of the crop and the foliar water potential . This last magnitude directly measures the water stress of the crop. By establishing these relationships, it is possible to calculate whether the plant is suffering from lack of water or not.

We have missed a few along the way. However, since this article was intended to be no more than an introduction to these indices, we will discuss the others in the not too distant future. In later posts we will go through the characteristics of each of them and their applications. Do not miss them!

Leave a Reply

Your email address will not be published. Required fields are marked *