Collecting Normalized Difference vegetation Index (NDVI) images is not a new practice on the farm. But how these images can be collected is changing as unmanned aerial vehicles (UAVs) gain popularity. That could raise some unique accuracy problems, says Agribotix co-founder Dan McKinnon.
"In theory, these NDVI images could be used for everything from prescribing fertilizer applications to estimating yields to identifying weed patches," he says. "However, the current state of the art is currently far behind the possibilities."
To understand why, it helps to understand how NDVI works, McKinnon says. NDVI was developed in the late 1970s by a NASA scientist using satellite imagery to compare relative reflectivity between wet and dry biomass. The scientist, Compton Tucker, developed a ratio of near infrared (NIR) reflectivity minus red reflectivity (VIS) over NIR plus VIS.
Clear as mud? Just understand that it’s a way to show contrast between plant and soil, and even between healthy plants and sickly plants, McKinnon says
Fast forward 30 years, and you no longer need a satellite orbiting the earth to collect NDVI imagery.
"It has become very (relative to 1977) inexpensive to modify a consumer camera to collect infrared bands and fly it aboard a small UAV," McKinnon says. "However, shadows cast by crops on a small scale and clouds on a larger scale dramatically affect these images, which is a problem that researchers working with satellite data never had to consider. We have observed this trend again and again – due to incident light variations, the NDVI image often returns crazy results."
Agribotix decided to use only part of the NDVI equation – NIR-VIS, also called the Difference Vegetation Index (DVI) – which is much less sensitive to differences in incident light than NDVI, McKinnon says. And image collection is only the beginning – the process isn’t complete until those images are ground-truthed.
Here’s one recent example, from a Colorado Dryland wheat field.
From the road, it’s difficult to see exact areas where the crop becomes thicker or sparser.
UAV-collected vegetative index imagery revealed big problem areas in the middle of the field that weren’t visible from the initial roadside inspection.