Landsat 8: California Central Valley

A picture is not worth a 1,000 words

NDVI: Normalized Difference Vegetation Index is one of many indices used in satellite, aerial and UAV imagery. Before we dive in with our thoughts on the matter, we thought it would be helpful to provide a bit of context on how NDVI is generated as well as explain how it is most commonly used.

To calculate NDVI values of a certain area, NIR (Near-Infrared Red) and some visible spectral bands are needed. The most common visible bands seen on imaging instruments are Red, Green and Blue. Plants absorb and reject different amounts of the electromagnetic spectrum based on their phenotype, general health and external stimuli. For example, a healthy plant rich with chlorophyll, absorbs shorter wavelengths of light; primarily Green and Red. At the same time, a plant rejects light at longer wavelengths beyond the visible spectrum like NIR. Therefore, a healthy plant will appear dark in Red due to light absorption and bright in NIR due to light reflection. The difference between NIR and Red is thus proportional to the overall plant health. We normalize this difference by sum of two values (NIR and Red) for equitable comparison. The equation for NDVI is often written as:

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

An unhealthy plant won’t absorb as much Red light as a healthy plant because of lack of chlorophyll/cell structure and thus its NDVI will be much less compared to more healthy vegetation with positive NDVI. Analytically, NDVI varies between -1 to +1, though realistic vegetation scenes have positive NDVI. By applying this ratio to crop vegetation such as corn, wheat, alfalfa, trees and grapes, general health can be inferred.

So with these fundamentals in mind of understanding how NDVI works we can assume crop health on a general scale of good, somewhere in the middle and bad. There is still more than meets the eye in order for our farmer to make better decisions. This is where the data fusion element comes in. In order to make NDVI useful one must know more than a color-scale photo that simply covers health. Items like crop variety, historical performance, weather in the region, growth stage, general management practices in ADDITION to using NDVI provide actionable information about a field’s bottom line: yield.

At Vinsight, we don’t just care about how green, yellow, or red your field is, we care about your output and fundamentally getting to a guaranteed prediction of yield with a low error rate. This allows the farmer to focus on the other critical elements of the business like making sure he will get the best price for crop on a per ton basis in order to make sure the investments spent on fertilizer, water, etc. will pay off.

Vinsight is continuing to provide yield estimates several months before harvest utilizing several other data sets in addition to NDVI imagery. Giving the farmer actionable insight into the field will allow for optimum growing for all crops and effectively change farming in the 21st century.