IEEE Future Directions
Related Articles
John Deere, one of the largest maker of agricultural machines, was criticised in the last years for the introduction of technologies in their agricultural machines because the farmers could no longer fix problems by themselves, as they have always used to do, but had to call in specialised assistance.
Now John Deere is planning to introduce even more technology but this time it is likely that farmers will appreciate that.
Modern agriculture has increased the yield by using insecticides to kill pest, herbicides to kill weeds and fertilisers to have faster growth. Both account for a large portion of agriculture today.
John Deere is working to cut this cost by equipping tractors with image recognition systems and AI to make precise spraying just where it is needed.
They have spent 305M$ to acquire BlueRiver Technology, a start up specialised in image recognition of weeds and are using AI algorithms to identify where, what and how much to spray to get rid of that weed. This leads to a reduction, so they claim, of 95% of the herbicide (with a corresponding decrease in cost) and is also improving the yield.
It should be noticed that some weeds have acquired a resistance to normal herbicides and have to be sprayed with more potent ones. The problem is that these are also affecting the crop, decreasing the yield. By minimising and localising with surgical precision the spray (known as “prevision agriculture”) the crop is not affected.
Machine vision and machine learning can be applied to all stages: seeding, tilling, weeding, watering and harvesting. In each one they can increase the effectiveness in a very significant way.
Now, I am no expert in agriculture and naively I considered this as an area “far” from information technologies, and even farther from AI. Yet, just few days ago I discovered, talking with a bank in Hungary, that they see an important role played by blockchains in the future of agriculture, because, they told me, the agriculture value chain is quite complex with big players at the edges (insecticides, herbicides, fertiliser companies on one side and canning distribution companies on the other) and many small players in between. The bank has no problem in providing funds to the big companies but has a tough time in assessing the risk of lending money to the small ones. By using blockchains they can trace the whole process and can evaluate the risk and hence provide funds to the small players too.
On a side note. I asked a friend of mine who works in the support of logistic chains an estimate of the number of lines of code involved in “managing” a banana, from its harvest to the distribution in a supermarket. He thought a little while and then said: given or taken one million lines, it is probably some 4 million lines of code. Not bad for a banana….
