Jun 10, 2022Digicrop views robotics, precision ag
Robotics and other technologies designed to benefit agriculture were at the forefront of Digicrop, the International Conference on Digital Technologies for Sustainable Crop Production. The virtual March 28-30 conference assembled researchers from a variety of disciplines who develop, use and review new technologies to improve the sustainability of crop production and breeding.
Precision agriculture uses intelligent automation to help farmers enhance crop production sustainably while reducing labor costs and inputs. Grower adoption, however, has remained slow and isn’t widespread, said Chandra Krintz, professor of computer science at University of California (UC) Santa Barbara and chief scientist and co-founder of AppScale Systems Inc.
While some large-scale industrial farming operations have begun to integrate AI and machine learning into their operations, many smaller operations haven’t. Existing solutions collect data using on-farm sensors, but move the data to the cloud for processing by different vendors and platforms. Because numerous vendors make the sensors, such services don’t work. The solutions and platforms don’t communicate with each other and are incompatible, Krintz said.
SmartFarm is an open-source computing system designed to provide farmers precision agriculture solutions. Krintz discussed how SmartFarm combines data from a variety of sensors, integrates recent advances in data analytics, machine learning and user interfaces that are compatible with systems available from public clouds. It eliminates the need for an IT staff to maintain the system.
“Many farmers have costly or poor internet connectivity, making those other solutions impossible to use,” she said. “Since intelligent automation needs to happen on-farm, moving the data to the cloud and back, it only causes delays and costs, while requiring farmers to share or give up ownership of their data.”
SmartFarm moves the cloud to the farm. In the familiar app store model, growers can download apps via edge cloud computer systems, which are low-cost and self-managing computing appliances.
SmartFarm systems are being investigated for use in different precision ag applications, including farm zone identification. The app identifies zones and reports them to growers for optimized management of those regions. In microclimate sensing for precision frost protection, physical modeling and sensor data can help predict the onset and duration of freezes, allowing growers to use the data to automatically control wind machines for air mixing or irrigation to warm the air and protect their plants against frost damage.
Root farming robotics
Robots are being deployed in central Europe to help sweet potato and seed potato growers counter labor shortages and costs. Alexander Langer, an industrial engineer and co-founder of Schmiede, a German firm that creates business models and robotic systems for construction and agriculture industries, discussed robotic automation in root crop farming.
In its second generation, Schmiede’s fully autonomous Harvey One uses stereo cameras mounted on the front of the machine as well as the intake belt to monitor crop row profiles and soil volume. Because every extra kilogram of soil growers must dig and work through affects fuel efficiencies, such monitoring is vital, Langer said.
The machines are being prepared for assisted harvesting, including automated weeding work. Langer said he’s working to produce the machinery in mass volume to lower costs and allow more grower adoption.
“What does the future of farming look like?” he said. “I don’t quite know, but the premise thus far has always been pushing economies of scale by creating bigger machinery, controlling your fleets through intelligent software, planning (and) analyzing to crank out that last bit of improvement. That may be viable for large-scale farming, but not for niches.”
A highly practical AI agricultural application involves building vision systems that can precisely recognize individual healthy plants, allowing growers to target treatments without damaging other normal instances.
Taeyeong Choi, a postdoctoral research associate in the Lincoln Agri-Robotics Centre at University of Lincoln, United Kingdom, studies AI computer vision and robotics for augmented intelligence in agriculture.
He discussed a system that collected 3,500 images of strawberries from his university using a mobile robot with a camera. The robot is being trained to detect normal and anomalous strawberries visually.
The research, however, showed that damaged fruit images were rare compared to normal fruit and that the robot would need to be better trained. Choi turned to Google Brain scientists, who solved a similar issue.
“What we expect of these agricultural robots is the capability of modeling the environment of the field precisely,” Choi said. “In this indoor strawberry farm, you can imagine using a mobile robot that can monitor qualities of individual berries using visual sensors. If the human farmers can learn from the robot which strawberries are unhealthy or damaged, they can perform special treatments and get a more accurate estimation.”
Other research involves real-time fruit monitoring and examining through a mobile robotics platform for strawberry inspection and harvesting within precision indoor farming systems (PIFS). A mobile robotics platform (MRP) is being developed to provide indoor farming operations the ability to monitor individual strawberry plants’ growing status and harvest ripe strawberries non-destructively.
The research, from Guoqiang Ren, a doctorate candidate in the Zhejiang University/University of Illinois at Urbana-Champaign Institute, saw an average 88% success rate. The research showed small fruit performed better than berries with elongated angles or enlarged body widths. “MPR has the ability to inspect and harvest strawberries within PIFS,” Ren said. “It can work smarter.”
The University of California and UC Cooperative Extension partnered with Digicrop 2022, which was organized by the University of Bonn, the German Cluster of Excellence “PhenoRob – Robotics and Phenotyping for Sustainable Crop Production” and the AI Institute for Next Generation Food Systems. Numerous presentations also covered crop breeding of grains and other commodities and digital and AI technologies to improve crop nutritional quality and abiotic stress tolerance.
– Doug Ohlemeier, assistant editor; Photo at top: Digicrop 2022 was a scientific event that targeted an audience working at the intersection of engineering, robotics, computer science, crop sciences, agricultural sciences, phenotyping and economics. Photo courtesy of Digicrop 2022.