History of the NDVI and the beginnings of biophysical parameters
The history of Precision Agriculture may have started back in 1972 when NASA launched Landsat 1. This technology allowed scientists to examine vegetation from space and determine how it’s developing based on analyzing colors on the spectral band.
In 1990 the GPS navigation system was developed and with it a new market for using satellite data. However, in the beginning, access was limited due to price and policy of the system operators. This was still before the time where public access to these systems really took off.
2015 – satellite data for agriculture
Satellite data usage in the area of agriculture really started as recently as 2015. That was when the European Copernicus program launched the Sentinel 2 satellites. Because Sentinel 2 operates across 12 different spectral bands, these new satellites were designed to be able to capture the spectral bands related to vegetation exceptionally well. Still, the vast majority of satellite services still are based on NDVI, even though this index only analyzes the red and near-infrared (NIR) bands of color. Because Sentinel 2 can access many more bands, there has been a lot of work in the last 3 years to develop new indices that can take advantage of the more powerful technology.
It’s along these lines that we’ve developed out Precision Farming service, which measures biophysical parameters to indicate the exact state of crops that are still growing. Several metrics are used to do this, including chlorophyll content, water content in the foliage, as well as vegetation density. To increase the accuracy, we calibrate our data using individual plots of land. During this process, we collect samples of the growing crops and perform laboratory analysis on them to make sure the predictions match the actual results. Next, we enter the results into a data model and the rest of the calibration of the satellite images in done through artificial intelligence.
From initial testing it quickly became apparent that the NDVI does not capture some key differences between different plots. As you can see form the figure, measuring biophysical parameters of crops is consistently more accurate. On top of that, satellite imagery could even suggest which crops would perform best on each individual plot.