On a recent summer morning on the plains of western Kansas, 50 farmers and new technologists gathered at the edge of a corn field to learn about using drones for farming.
Sponsored by the Kansas Agriculture Research and Technology Association, attendees ranged from growers practicing precision agriculture to drone manufacturers and agricultural-services companies. All see the potential for drone technology to help usher in the next phase of the precision-agricultural revolution, adding new tools to help growers maximize yields and save money.
At the workshop, growers got a taste of available drone capabilities, with technology vendors demonstrating their remote-sensing systems. A short while later, they examined near-real-time results with attendees over lunch. Growers saw firsthand the primary benefit of drone technology for agriculture: the ability to quickly and easily image a field, and in a few short hours have a map in-hand showing relative crop condition.
Improving Previous Methods
Basic crop scouting from drones provides a major improvement over how field surveys previously were done, and it's the primary driver for drone-technology adoption. Crop scouts traditionally were employed to walk farm fields to monitor crop condition. This is hot, time-consuming work, and crop scouts have difficulty visiting even a small percentage of a whole field, easily missing problem areas.
With the advent of drones for agriculture, crop scouts now can arrive at a field, launch a drone, collect near-infrared imagery over the entire field, upload imagery to a cloud data-processing service and, in a matter of hours, receive whole-field maps showing relative crop condition. The crop scout now can easily visit specific target areas within the field that need assessment via boots on the ground.
The Promise of Remote Sensing for Agriculture
Aerial and satellite imaging have shown potential for commercial application in agriculture for decades, but the promise hasn’t been fully realized. Long satellite repeat cycles, expensive imagery acquisition and difficult-to-use software tools all conspired to hinder the adoption of remote sensing in agriculture, but that's all changing. With the advent of low-cost, easy-to-deploy drones and cloud computing, the promise of remote sensing for applications in agriculture finally is being realized.
New companies such as Agribotix, who I work for, are delivering new capabilities to a global agriculture market enabled by the remarkable pace of innovation seen across technologies ranging from cloud computing to open-source software to advances in robotics and aerial systems. When combined, they become a force-multiplier, enabling companies to rapidly innovate and deliver new products to market at affordable prices.
Looking at the speed of innovation, consider the development of multi-rotor drone platforms for agriculture applications. Multi-rotors have several advantages over fixed-wing aircraft:
- Precision takeoff and landing, which can often be a challenge for fixed-wing drones.
- Improved stability in heavy winds.
- More-precise flights where flight lines are straight and transitions between flight lines are smooth.
But as recently as a year ago, conventional wisdom held that multi-rotors didn't have the range necessary for covering large farm fields. However, with the rapid pace of innovation in battery technology and avionics and design, the endurance of multi-rotor platforms now is comparable to fixed-wing aircraft. Multi-rotors now can easily fly 160 acres, the size of a quarter-section central-pivot irrigation field, in one flight with one set of batteries.
New applications are emerging that are only possible with the unique data-collection capabilities offered by drones: low-altitude flight and the ability to position sensors within meters of a target. For agriculture, growers and crop consultants can fly a multi-rotor drone at low altitude over a farm field and image growing crops at unprecedented spatial resolution. Combining such high-resolution imaging data with cloud computing and cutting-edge machine-vision algorithms, new applications offering real economic return are possible, and they're getting easier and less expensive all the time.
A great example of a new drone-powered product for agriculture is crop-stand count: counting individual plants early in the growing season so growers can make informed management decisions. Early in the 2015 North American growing season, much of the Corn Belt experienced cold and rainy weather, conditions that can impact corn-seed germination and emergence. Having timely information on corn-stand count over a recently planted field can be invaluable for farm managers, providing answers to the following critical questions:
- Â Should I replant a portion of the field?
- Â Do I need to modify the planter setup?
- Â Do I have skips or double plants?
- Â Did I achieve population planting goals?
With drones and the collection of high-resolution imagery that can resolve individual, recently germinated corn plants, new classes of computer-vision algorithms can be brought to bear. Computer systems can be trained to recognize individual corn plants based on how they look. This is unprecedented in the context of traditional remote sensing, where the focus has often been on multispectral and hyperspectral single-pixel analysis for materials identification, and where high spatial resolution can be measured in pixels that are tens of centimeters on a side.
With the advent of drone-powered remote-sensing applications, these are exciting times for Earth imaging across a host of commercial industries. For agriculture, and in concert with cloud computing and new classes of machine-vision algorithms, drones are reducing growers' costs and increasing their yields, while more, new and exciting drone applications are waiting to be discovered.
Thomas Harris is vice president of Products and Data Services at Agribotix; e-mail: firstname.lastname@example.org.