While satellite imaging of farm land and unmanned aerial surveillance of crops capture most of the attention in remote sensing, a young analytics firm and Valley Irrigation have been collaborating over the last couple of years to bring crop monitoring a little closer to the ground with sensors mounted on center pivot sprinklers.

Prospera Ag made headlines of its own in 2019 with pivot-mounted sensors to study anomalies found in agricultural aerial data and unique methods of controlling variables in lighting when monitoring crops from the moving irrigation spans. 

Daniel Koppel, Prospera’s CEO, says his company’s experience showed good opportunities from a pivot-based sensor system to research specific field conditions and invited Valley Irrigation to become part of an experiment to develop the technology.

“By being mounted on the pivot, only a few meters from the plant, using an imager designed for work at nearly the crop level, we found we could do more sophisticated analysis than with flybys with aircraft,” he explains. “Valley has access to 10,000 soil probes across the country, so with weather data and our sensors, we saw a new approach to decision-making.”

By combining on-pivot information with aerial imagery, soil probe information, soil type records and weather data, Prospera and Valley are using machine learning tools to interpret what the crop is showing.

“One piece of data is not enough,” Koeppel says. “With multiple data inputs you get a better sense of the puzzle, and with artificial intelligence (AI) we can digest multiple layers of data to alert farmers of plugged nozzles, early disease on-set or the presence of pests.”

The on-pivot sensors eliminate many variables encountered while shooting in sunlight — sun angle, time of day, overcast skies — by photographing at night with a flash calibrated to conditions.

The companies spent 2019 working on verifying the results of their sensing tools and artificial intelligence-based crop monitoring and detection service, in Washington, Texas, Nebraska and Idaho.

By late April this year, they announced they were quadrupling the coverage area of their work in the U.S. and voiced their confidence in the system’s ability to deliver “actionable insights” to help growers reduce inputs and improve yields.

“We worked closely with growers to incorporate our feedback and develop the next generation of the service for this broader commercial roll-out,” says Valley Irrigation’s vice president of business development, Darren Siekman. “The system can detect irrigation machine issues, over- or under-watered areas and other issues. This is going to give growers the opportunity to improve their bottom line as the Valley Insight system alerts them to potential problems.”

“Collecting high-quality data for improved crop productivity has always been cost-prohibitive,” Koeppel explains. “By combining our industry-leading computer vision and AI technology with the leading tech-enabled infrastructure of Valley Irrigation, we’re reducing the costs of data acquisition, increasing the speed of deployment of new technologies, and delivering real value to growers.”

Koeppel says 2020’s experience on cooperator farms across the U.S. will develop data sets to fine-tune the system’s abilities for a 2021 rollout to complement existing aerial and satellite imagery. 

Sans Sensors

Meanwhile, in other irrigation-related developments, CropMetrics, now CropX, is perfecting its Virtual Predictor tool, launched last summer as a “virtual soil probe” that relies on modeling tools combined with different data streams and agronomist insights to make crop decisions.

“Virtual Predictor feeds information on crop and soil conditions, weather and evapotranspiration into a powerful model that allows us to make soil water forecasts,” says John Gates, chief science officer for CropMetrics. “It gives you an estimate of your soil moisture and soil water status over the next 7 days.

“Agronomist will have access to the information and will leverage that into the recommendations they write.”

A number of irrigators are using this and similar modeling products in the High Plains of Texas, Oklahoma and Kansas to operate without the expense and field disturbance of soil moisture monitors. In some cases the expense of the subscription and programs required are fully offset by the lack of moisture probe rental and installation.