Full-value or half-value from precision ag efforts?

Kristina Polziehn visits an irrigated cotton field in Australia on her Nuffield tour.  In this photo she is with husband  Glenn Wormald who is also a Nuffield Scholar from Australia in 2017.  |  Nuffield photo

Producers collecting multiple layers of field data complain they aren’t milking the data for all it’s worth.

“I’m not even getting half the value out of these images.”

Kristina Polziehn hears the similar complaints. She says there definitely is an obstacle to applying the data and images in real-world precision agriculture practices. That obstacle is the lack of human touch.

Polziehn made the observation in her recently released Nuffield Report titled Remote Sensing in Crop Production: Applications and Value. Polziehn, who owns and operates an agronomy business near Edmonton, produced the report sponsored by a Nuffield $15,000 scholarship and the Alberta Wheat Commission.

The report focused on extracting better value from the imagery and maps employed by the precision ag industry.

“A lot of the data we’re capturing can be utilized in precision ag once it’s all put together to tell a complete story about what’s happening on a field. However, this does require specialized expertise to bring it together.

“We getting all kinds of satellite imagery and aircraft and drone aerial imagery. That’s good, but those images need to be combined with your soil test maps, EC maps, yield maps, topography data and other layers. Without ground proofing, you’ll never get your full return on satellite and aerial imagery.”

Polziehn says there’s nothing new about layering maps. The challenge is how to use those layers, and that’s what’s missing in the effort to extract maximum value from imagery.

She says there’s a trend on the part of some bigger companies to automate the process of interpreting data. She says the trend is to take all the layers of information and just stack them together, a process that eliminates the personal touch. She emphasizes that the process still needs a human to deal with factors algorithms cannot see.

“An agronomist needs to ground truth the interpretation the computer spits out. Does the automated report match what you see in the field? That’s the only way you’ll maximize the value of your imagery. It’s the only way a producer can say he’s getting all his money’s worth, not just half his money’s worth.

“For example, if you look at a satellite NDVI image of a field and you see two areas that are high production. What the image doesn’t tell you is that one area is producing a lot of grain. The other high production area is overrun by herbicide-resistant weeds. You need ground truthing to discover these things and to get full value from the images.

“We’re in an industry where people tend to specialize in the agronomy side or specialize in the precision ag side of things. There’s few people who bridges those two worlds.”

In the first year of her study, Polziehn used a fixed wing UAV. But that didn’t last long.

“I did the drone thing. I bought a fixed wing. Here’s what I found. By the time you drive to the field, weather conditions might have changed so you can’t fly. So you have to come back the next day hoping conditions were better. By the time I analyzed the money, the hours and the setup for just a quarter section, I decided this was ridiculous. I decided to hire a plane.

“With a Cessna, I can easily fly 5,000 acres in one day, and get better quality imagery. You’ll never do that with a drone. Breeders for example demand very high resolution. I didn’t hire the special photography Cessna. I just shot out the door. If you’re flying big projects, it’s definitely more feasible to use the plane. If somebody has 600 or 700 acres they’re concerned with, I tell them ‘sure we can fly that in an hour’.

“If I have a client who needs really high resolution for their project, I can get 10 centimetres or less with a manned aircraft. If you need two centimetres, I can even do that. I’d need to get different cameras and lenses, but that’s the kind of thing you can do with a manned aircraft. If you need one-inch, I can do that.”

Polziehn says the Nuffield Scholarship allowed her to travel to numerous countries around the world to learn how growers and agronomists can use remote sensing imagery in support of other precision agriculture tools and practices.

Nuffield Canada awards several $15,000 scholarships yearly.

Excerpts from Polziehn’s Nuffield Report

The advancement in sensor technology to measure spectral emittance and reflectance has not only improved our understanding of agronomic factors but also improved its application in crop management.

This report reviewed different examples of applications of remote sensing in crop production, through global travel.

Under various cropping systems, remote sensing data has been used across many crops to monitor plant growth and detect plant stresses as they relate to water, nutrient and pest management.

In addition, progress continues in developing remote-sensing tools for monitoring crop development, screening new crop varieties and yield modelling.

The report further examined the value remote sensing data has for farmers and agronomists. Remote sensing can provide important information for crop management decisions but can also assist in determining management zones for VR technologies.

One obstacle for remote-sensing imagery adoption is the direct economic value it provides, as many of the benefits are realized through a combination of other precision agriculture technologies.

Many aerial imagery data applications require additional analysis and interpretation to become actionable and usable, which can translate to additional costs and outside expertise for usable applications.

Similarly, further value of the remote-sensing imagery comes when it is combined with additional data pertaining to agronomic practices, weather and climate, soil and plant development and rolled into a comprehensive, functional tool.

In Canada, more focus is needed to develop and expand these tools through present crop modelling initiatives and international collaboration.

Current adoption of precision agriculture technologies such as the use of remote-sensing data remains slow, despite a long, embedded history in crop production.

Progress to move forward our adoption of these technologies will require more investment of money and time into precision agriculture workshops, technology transfer programs, precision agriculture university degree programs and high-level precision agriculture research pertinent to Canada.

The full report is available at www.nuffield.ca/scholars-reports.

About the author


Stories from our other publications