Drones don’t just take pictures, they capture a wealth of data about your crops’ health. But how do you know which data is best to use? In this post, we will compare RGB and NIR imagery and explain which we prefer to help you keep your fields healthy and improve yields.
Let’s start at the beginning. What types of drone cameras are used for identifying crop variability?
The human eye is sensitive to red, green, and blue (RGB) bands of light. Most standard drones come with cameras that capture the same RGB bands so the images they produce recreate almost exactly what our eyes see. If you are just starting out using drones on your farm, it’s OK to start with RGB cameras since they are fairly cheap and most consumer drones like the DJI™ Phantom and DJI™ Inspire systems come equipped with them off the shelf. RGB cameras are good for creating orthomosaic maps that show your entire field at once, and they can capture aerial videos.
How RGB Sensors Can Help:
Faster Field Scouting
Traditional crop scouting involves gas-guzzling ATVs, trucks, or tractors; driving through your fragile fields; and walking to manually observe plants and find problems. With an RGB camera on your drone, you can see your entire field in one place by processing aerial imagery into an orthomosaic map, quickly make observations, enter the geodata into your GPS, and drive right into the problem area without damaging your entire field.
Drones empower you to fly and capture imagery or live video from a higher vantage point so you can inspect assets safely. Use drones with RGB sensors to inspect assets such as grain bins, barns, or other buildings and equipment more efficiently and effectively.
Approximate Crop Health
Some farmers have used an index that works with RGB sensors called Visible Atmospherically Resistant Index (VARI) to detect areas of crop stress. The VARI algorithm uses some color correction to minimize reflectance, scattering, and other atmospheric effects to better estimate the fraction of healthy vegetation in an area. Effectively, it exaggerates color and shows how green the plant is in comparison to others so you can approximate plant health and vigor. VARI is not a replacement for the Normalized Difference Vegetation Index (NDVI), but it can be useful if you only have an RGB camera which cannot process NDVI imagery.
Near-infrared (NIR) sensors capture light invisible to the human eye. There are two types: Multispectral cameras which capture Red, Green, and NIR light and Modified RGB Cameras that are enhanced to capture NIR as well as RGB light. NIR sensors are more expensive than RGB sensors, but let users do much more.
How NIR Sensors Can Help:
You can do everything with an NIR sensor that you can do with an RGB sensors and more.
Process NDVI Imagery
Normalized Difference Vegetation Index (NDVI) is one of the best indices for evaluating crop health and NIR sensors are necessary to measure it. NDVI mathematically compares red and NIR light signals to help differentiate plant from non-plant and healthy plant from sick plant. NIR sensors and drone software make it simple for agronomists and farmers to benefit from NDVI by creating maps that convert the ratios of invisible NIR light into colors humans can see and quickly evaluate.
Gather Accurate Crop Health Data Over Time
VARI can vary. Puns aside, VARI is not normalized for variations in sunlight, cloud cover, shading, etc., which means flying a field from one day to the next could reveal very different results. Even if you fly half your field at dawn and the other half at high noon, the results might not be comparable.
Because it incorporates NIR, NDVI by contrast is normalized for such variations, making it a better tool to compare the health of your crops over time, whether it be week to week, month to month, or even season to season. Comparison is an important part of agronomists work in order to improve crop yields, so NDVI is a necessary tool for the job.
Identify Problems Sooner
Plants show stress in NIR sooner than the visible light spectrum which means that you can detect even the slightest crop stress, days or up to weeks before you’d be able to with the naked eye or an RGB sensor. NDVI mapping helps you detect problems sooner so you can take action to improve crop health before it becomes detrimental to your yields.
Create Variable Perscriptions Maps
NIR sensors create NDVI maps that you can upload to agriculture software like Ag Leader® SMS™ or agX® SST to identify variability in your fields. Your agronomist uses these software tools to generate prescription maps of your fields, showing you exactly where to seed and spray with variable rate applications that save you money and resources, while improving crop yields.
How Botlink AssessES Crop Health
Botlink prefers NIR sensors because we want to give you the best data you can get and provide you with the most accurate indication of plant health every time. NIR sensors create NDVI maps that detect problems sooner and help you make more accurate comparisons throughout the year or season to season.
Botlink doesn’t use VARI for crop health because we don’t believe it adds significant value beyond what a standard RGB orthomosaic shows. In some cases, VARI data can be even harder to evaluate than standard RGB, creating more problems than solutions.
Debunking the Myth of Synthetic NDVI
While searching across the web, you have probably seen the terms “synthetic NDVI” or “false-NDVI” which refer to an NDVI algorithm applied to RGB imagery. It’s called false-NDVI because it doesn’t collect NIR light that NDVI normally requires, and instead uses RGB images to approximate NDVI. We don’t recommend using false-NDVI as it reports inaccurate data that can be detrimental to your crops.
If someone points you towards false-NDVI, we suggest looking at the other options mentioned above. The VARI algorithm is a more useful and accurate way to measure plant health using RGB imagery. If you can get an NIR sensor, we always recommend true NDVI as your best option.
While drones make field scouting faster and more effective, they can’t replace it all together. Using the algorithms to measure the plants shows the variability, but it can’t tell what causes that variability. That’s why no matter what method or calculation you use, it’s always important to ground truth and take samples. Once the variability is found, put on some boots, go right to the source of the problem, and determine the best course of action.