Robotics are touted as a natural evolution of the automation of agriculture...
Agricultural automation on the farm has advanced in fits and starts. Drones have mostly turned out to be expensive toys, yielding a return on investment only for large operations. Other forms of automation may offer a faster ROI for some repetitive, labor-intensive chores like milking a herd of Holsteins or boosting crop yields.
A market researcher argues that agriculture automation may have turned the corner with advances in two key core technologies: machine vision based on convolutional neural networks (CNN) and “autonomous mobility.”
The key advocate for this position is IDTechEx, which argues robots, machine vision and other forms of automation are, unlike Bob Dylan, going to work on Maggie’s farm.
“Agricultural robots and AI-based machine vision represent the natural evolution of agricultural machines and tools,” the U.K-based market tracker asserts.
While short on forecast numbers, IDTechEx does identify 15 robot types and agricultural sectors ripe for automation. Those range from autonomous spraying vehicles and tractors to time-sensitive robotic fruit harvesting.
CNN-based machine vision is touted as enabling robots to distinguish among individual plants and weeds. The payoff would be precision weeding that would reduce use of herbicides while boosting yields. Once trained, “ultraprecision robots” might be able to analyze the condition of individual plants. Row vegetables are a prime candidate.
Indeed, researchers have proposed using CNN-based machine vision and pattern recognition for agriculture applications like plant seedling classification.
“Previous machine vision technologies developed for selective weeding have faced the challenge of reliable and accurate weed detection,” researchers noted in a 2018 paper. They found that “CNN-driven seedling classification applications when used in farming automation has the potential to optimize crop yield and improve productivity and efficiency when designed appropriately.”
Farther afield (sorry), robotic tractors equipped with cameras and GPS could plow fields or apply fertilizers. The key market for these forms of automation is large agricultural operations, and IDTechEx said the latest versions can travel at speeds up to 12 mph while processing images at 20 frame per second while maintaining 2-inch resolution.
Over the next two decades, the market analyst forecasts that adoption of technologies such as “tractor guidance,” fully autonomous vehicles and robotic harvesters will more than double.
The agricultural “hardware and machine vision evolution will inevitably continue, just as with all other agricultural tools and vehicles,” the analyst notes. “We are still at the beginning. The deployed fleet sizes worldwide are small, but this is about to dramatically change.”
Therefore, it appears you we’ll soon be able to keep the robots down on the farm.