The Intelligence System for the Ag Drone
Small, low-cost unmanned aerial systems, or “drones” as they have become more commonly known, are providing new platforms for observation that can and will transform several industries around the world. You’ve probably seen stories to this effect in the press over the past couple of years.
Many ask what drones can provide that manned aircraft or satellites cannot. From a remote sensing point of view, the low- and slow-flying drone enables extremely high resolution and accurate measurements that have not otherwise been achievable at an acceptable cost. From an agriculture management point of view, this means on-demand access to valuable information that has never before been available.
The drone itself however, is just simply a platform for observation. The value that a drone ultimately provides is tied to the value of new information it provides to its user through the technology chain of sensors, processors, storage devices, networks, software, and user interfaces that we collectively refer to as the drone’s intelligence system.
SLANTRANGE is the drone’s intelligence system. And the following is a short introduction to the technical principles and the approach we take toward offering the best possible information, or “intelligence”, to support your needs.
What Is Spectral Imaging?
Even if the term is new to you, you are familiar with spectral imaging. It’s a technique that exploits the wavelength-dependence of light reflectivity from an object to infer new information about the nature of that object. More simply put, we can learn things about an object from its color. In fact we’re all familiar with spectral imaging when it comes to vegetation – we know that as fall arrives the green hues of summer give way to oranges and reds.
But have you considered why those colors change? As a plant’s health and maturity evolve throughout the season, its chemistry and cellular structure evolve as well. Those changes are evident in the sunlight it absorbs, transmits, or reflects. In fact a leaf’s color is telling you about what’s happening with its biological processes. During the spring and summer, in a plant’s most vigorous stages of growth and maturity, chlorophyll and other accessory pigments soak up the blue and red portions of the sunlight that strikes its leaves and reject the green portion. That’s why healthy looking plants are deep green in color. In the fall, as plants senesce, photosynthetic activity decreases and more of the red portion of the spectrum is reflected.
The basic principle underlying SLANTRANGE’s sensor systems is the same that you’ve been using your whole life with your own eyes.
So Much More Than Your Eye Can See
If your eyes were more sensitive to deeper shades of red, or had a better ability to resolve differences in color, there’s a lot of new information you could take from looking at the world. The chart on the left shows the amount of sunlight a leaf reflects across the spectrum of colors (wavelength). A healthy plant reflects about 10-15% of the light that strikes it at 550 nanometers wavelength (commonly known as green), but actually reflects substantially more sunlight above 700 nanometers, in the infrared portion of the spectrum that is invisible to the human eye. The rainbow-colored bar at the bottom of the chart shows the region of the spectrum that humans can perceive.
A stressed plant has a different “signature”. It’s reflectance profile shifts as the balance of chlorophylls, carotenoids, water content, and other pigments changes in the plant. Furthermore, different types of stress result in slightly different signatures. While invisible to you and to a normal camera, the signals are there and can be measured with SLANTRANGE sensor systems.
The Importance of Calibration & Sensitivity
If your objective is to measure the continuing status of a plant’s health using a spectral imaging approach like that described above, there are two critical elements to the design of a sensor system that must be considered.
First, the sunlight we receive on the ground is constantly changing. At noon, the light is a bright white. In the mornings and evenings, deeper oranges and reds dominate. Smog, haze, humidity, clouds – all have an effect on the color and intensity of sunlight.
So what happens if we photograph a plant at noon and then later in the day? Or if a cloud layer comes in? The two photos will look different, not because the plant has changed, but because the mix of sunlight changed.
Second, the signals that indicate a change in a plant’s health condition are initially very small. In spectral terms, a changing condition may be evidenced by a small “spike” in the spectrum only tens of nanometers wide. Over time, as the plant deteriorates, the width and amplitude of that spectral spike will increase until eventually it becomes perceptible to the human eye.
A system designed to make meaningful spectral measurements of plant health therefore requires:
- A mechanism to correct for changing lighting conditions.
- Sufficient resolution and sensitivity to isolate important plant signatures from background clutter.
SLANTRANGE’s sensor systems, like the one pictured here, are designed specifically to meet these requirements. SLANTRANGE systems simultaneously measure sunlight and the crop signature in narrow bands of the spectrum that are critical to quantifying plant health – so crop health can be tracked over time.
Edge Computing, Speed, and Scalability
Crop health information is perishable, particularly in cases where a failing condition exists that demands immediate action. A delay of only a few hours can result in tens of thousands of dollars in lost yields.
The challenge for airborne agronomy is to produce useful information quickly and locally so that decisions can be made and action taken. Time spent hauling data back to the office for analysis or uploading it to a cloud server for overnight processing not only costs money but also eats away at the value of the information. Moreover, most farmland isn’t blessed with high bandwidth network connectivity or high capacity computing platforms, especially outside of the US.
Our approach is different. First, our processing algorithms minimize the amount of image overlap required to produce results, cutting your data collection time down by up to 4x compared to similar narrow-band multispectral systems. Second, our systems produce results immediately at the point of collection with no need for network connectivity, high-end computer systems, or trained operators.
We’re putting valuable information into your hands while it is still valuable, anywhere in the world.
Vital New Information for Agriculture
What information is most valuable? For the producer, the agronomist, the supplier, the insurer, the trader?
SLANTRANGE’s measurement systems are quantifying crop development in ways never before possible with satellite or manned aircraft systems. With spatial resolutions down to a couple of centimeters, we can now rapidly isolate every plant in a field for analysis – and that enables some truly unique capabilities:
- Crop population statistics – Plant density and size information to support decisions for replanting and baselines for insurance.
- Weed population statistics – Detect individual weeds down to a few centimeters for targeted herbicide applications.
- Stress conditions – Pest infestations, nutrient deficiencies, and dehydration conditions can all be detected and zoned for treatment.
- Canopy closure – Quantify the percentage of a field with a closed canopy to support yield estimations and substantiate crop damage insurance claims.
- Yield potential – An intelligent combination of several other metrics to estimate potential yield at harvest.
Introducing the Time Dimension
Yield monitors and GPS-guided implements were a revolution when they were introduced – suddenly producers were able to visualize how yields developed in two dimensions across their fields. Those technologies however, only provide postscripts to the events of the season at the time of harvest, too late for action.
Because SLANTRANGE systems are inherently self-calibrated for changing environmental conditions, in-season measures of plant stress can be compared over time and correlated to other variables to determine the leading factors driving yield. Trends can be identified, better forecasts can be made.
What SLANTRANGE Means For You
SLANTRANGE offers a complete measurement and analytics system, tuned to provide the information you need, when you need it, and under the conditions in which it needs to be employed. That means:
- Trust that your results are a function of plant conditions, not weather conditions, time of day, or season
- Earlier indications of emerging health conditions that need attention
- Faster results
- A record of field growth throughout the season, year-over-year, to correlate against your inputs, to understand better what’s working or not working with your management plan
Have you ever asked yourself what you would have done differently if you had the data in hand in June to show you what your yield map would look like in November? Now you can.