Farmers, agricultural retailers, equipment manufactures, and researchers all are faced with a multitude of fast changing dynamics in the agricultural industry - technology and the associated data are two of the biggest. The development of new technology as well as its adoption and implementation are going to be keys to future success for all those involved in the industry.
“The question is not if a farmer is going to adopt technology but how and when they are going to adopt technology,” said Dr. Scott Shearer, Chair of the Department of Food, Agricultural and Biological Engineering at The Ohio State University.
Companies supplying new farm equipment and technologies will need to work closer with those that supply agricultural inputs to develop new seed, chemical and fertilizer technologies. In the same way, all of those companies will need to work together with the livestock producers that feed the crops produced.
“Everyone needs to look for efficiencies,” Shearer said. “The equipment manufacturers need to realize that a lot of grain is marketed through livestock both as grain for feed and also the grain byproducts from ethanol and soy biodiesel production.”
Agriculture is often segmented into either crop production or livestock production. Inputs are divided into those items used up each year (seed, chemicals, and fertilizer) and longer-term assets such as equipment and technology. In reality, the agricultural industry should be considered a seamless system, extending from the inputs used all the way to the end consumer.
“We have already started to see this transition as chemical companies now own many seed companies,” Shearer said.
Equipment companies, as well as the livestock and biofuels industries, are not far behind, according to Shearer.
“The industry as a whole is paying more attention to environmental factors like weather, soil health, water quality, and developing strategies for more efficient management of inputs as a result,” he said. “There are nearly 200 million acres of crop production in the Midwest. Farmers are constantly looking for ways to increase the use efficiency of the inputs needed to produce crops. As machine learning and artificial intelligence (A.I.) expand to include the management of crop inputs, while factoring in environmental conditions, automation will reshape cropping systems.”
“Until recently, most A.I. work in agriculture was performed in greenhouses or at universities. Research was conducted with a limited number of variables based on the environment. In the past, digital and infrared photo equipment was limited to about 12 features to capture and measure. Now, up to 10,000 features can be collected and measured. The concept of ‘training neutral networks’ has become a reality in agriculture as we now have machines with the computer power to perform the data collection and processing.”
Training neutral networks involves using a series of the data points collected along with modeling to create a mapping of inputs to outputs. The training (machine learning) process is solved using an optimization algorithm that searches through a set of possible values to determine a more efficient model that will result in the optimized performance. A.I. incorporates training neutral networks that combine forecasting, data validation, and managing risk.
“The first use of this technology will be with spray application. This will be the litmus test for automation,” Shearer said. “Aerial imagery will be used to classify numerous aspects of the plant canopy.”
The technology currently allows satellite imagery or a drone to collect aerial color and infrared imagery indicating areas of plant health concern. Those locations can be further investigated with a drone to “take a closer look” and diagnose plant stress causes (disease, insects, fertility) by leaf tissue color. Equipment has even been developed for a drone to lower a camera into the canopy for additional photos of the plant below the canopy and collect tissue samples that are retrieved for further analysis.
This level of technology generates extensive data to be managed.
“There is a real need for agriculture subject matter experts to help the data scientists analyze and access the data,” Shearer said. “There is a whole lexicon of terminologies that are needed to bridge the gap between the scientists and the agricultural experts. Two other influences in this process will involve both the data ownership component and data security and privacy. Will we need the equivalent of HIPA for agriculture like currently exists in the health industry?”
While most companies say the farmer owns the data, that can be misleading once the data hits the cloud.
“What is often overlooked is the fine print of the cloud storage locations. This is often found in the user agreements and is for the company protection. By clicking on ‘I accept’ basically, the farmer is licensing their data over to the companies,” he said. “Not every technology will return the same value to every farm. A farmer needs to find the technology that will do what they want, and then implement it,” Shearer said. “Most older technology is now considered standard technology on farm equipment. There is a continuum of adoption of technology in the ag industry. There are those on the cutting edge, or as some say ‘bleeding edge,’ that are the first to adopt technology before it is completely proven, and they are the ones that sometimes get hurt by it. The farmers who make money implementing technology are usually the ones near the middle.”
Technologies are tools in the toolbox for farmers. They need to prioritize the things that are important for their unique operations, and then find and utilize those technologies to ensure long term profitability.”
By: Dusty Sonnenberg, CCA, Ohio Field Leader
Dusty Sonnenberg is a farmer and certified crop advisor. He provides agronomic content for the Ohio Field Leader as well as Ohio’s County Journal. Dusty is an active member of his community and owns and operates Sonnenberg Farms, Jay Calf Ranch and Tri-State RTK Network, LLC.
