As my wife and I drove down a rural Iowa highway last week, we began speculating about the next revolution in field management. Currently, tractors blindly and mechanically groom rows of crops spaced wide enough to accommodate their massive tires. Large quantities of herbicides and insecticides are broadcast, leaching into streams and aquifers. In turn, specially-bred seeds, resistant to these chemicals must be purchased as a part of a proprietary program.
Instead, I imagined swarms of spider-shaped robots with travel legs long enough to keep their body suspended above the crops. Work arms, tipped with cameras and tools, could maneuver to any spot.
- Plants could be grown in an efficient honeycomb pattern, wasting less space.
- A database of every plant, with its progress, could be maintained.
- Instead of making care decisions on a whole-field basis, adaptive algorithms could adjust interventions for increasingly small areas.
- Micro-doses of fertilizer nutrients could be injected under the surface, sufficient for each plant.
- Individual weed plants could be identified and selectively uprooted.
- Individual bugs could be identified and selectively destroyed.
- Individual bugs could be harvested using a suction device. Some bugs contain valuable chemicals such as dyes or pharmaceutical components.
- Crop plants could be automatically thinned or even transplanted to more-sparse areas.
- Robots could work continuously, even at night, returning to an energy source to recharge or exchange batteries.
- Some labor-intensive jobs, such as harvesting strawberries, currently require lots of labor for brief periods of time. Machines would be easier to store and transport from place to place than people.
The next day, I discovered that researchers at Leibniz University in Germany are exploring how to use lasers to kill weeds. Cameras feeding pattern-recognition software can identify multiple weed plants and distinguish them from the crop plants. Tunable lasers, aimed at the most vulnerable areas, can kill weeds. In the process, the German researchers found that lesser intensities of laser light actually stimulated weed growth. This raises the potential of non-chemical stimulation of crop growth.
Researchers in Israel are developing multispectral sensors for identifying fruits and vegetables along with their ripeness. They are already able to correctly identify 80-85 percent of fruit on a plant. They are also designing grasping tools that can remove individual pieces without damaging them.
This is a field (pun intended) to keep your eye on. The core issue of agricultural productivity has always been the limitations of manpower. Perhaps it is time to look away from ever-larger mega-machines. These are becoming highly-automated themselves anyway. The next step is to teach smaller highly-automated machines to perform the tedious judgment-intense precision farm work that we can no longer afford to do in person.
©2012, David Satterlee