Machine Learning Applied to Agriculture

By Johnny C.

Technology is always changing and growing. In the modern day, advancements in technology come at exponential rates. One of the fastest ways to get ahead in an industry is by investing in cutting-edge technology. Agriculture is a huge part of the economy and is very important in sustaining countries. A more recent development in agriculture industry, is machine learning. Machine learning is a type of artificial intelligence that provides computers with the ability to learn without being explicitly programmed (Rouse, 2016). An example of this is, is an advancement made by biologist David Hughes and epidemiologist Marcel Salathe. “They fed a computer more than 50,000 images, and by learning on its own, the program can correctly identify 99.35 percent of the new images they throw at it.” This is a huge advantage for farmers, allowing them to understand what is affecting the crops, and finding faster ways to cure them. Machine learning can also help discover the best times and locations for crops, maximizing the yield. Although there are limits and challenges to machine learning, it is capable of pushing the boundaries in agriculture.

There are a plethora of benefits pertaining to machine learning. Farmers are able to use computers to record patterns of their crop growth to help figure out the best locations or times. It helps alleviate time that farmers would be using to raise the crops, allowing them to focus on other aspects of farming. By letting computers handle the more trivial matters such as pinpointing certain plant characteristics, farmers can take advantage and allocate their resources to maintaining the farm. Another benefit is by allowing a machine, such as a tractor, photograph the crop identifying if certain plants don’t belong. A big problem farmers have is taking care of weeds, and they typically deal with it by spraying pesticides which may potentially damage the important crops.

The tools and software used ranges from satellites, computers and cameras. Programmers then create algorithms that help organize all the photos and scans taken, and then use them to aid the farmers. “US government run satellite programs such as LandSat or MODIS can take an image of the entire globe at 20 to 30 meter resolutions once a week. Now new nanosatellite constellations, can take snapshots of the entire globe at 3 to 5 meter resolutions per day” (Brokaw, 2016). One example of machine learning is Mark Johnson’s startup company, Descartes Labs. One of the biggest crop in America is corn. Johnson is trying to beat the United States Department of Agriculture(USDA), at predicting corn yields. The USDA deploys workers to survey and scout out farms during the busiest seasons, whereas Descartes Labs work from their office. By using new nanosatellite constellations, Descartes Labs is able to work without setting any foot outside, and accurately forecast corn yield predictions. Another application is from a company called Blue River Technology. They have created a machine called the LettuceBot. With this machine learning tractor, Blue River hopes to reduce chemical use by accurately pinpointing which plants to spray. The LettuceBot photographs 5,000 plants a minute and is able to discern between lettuce or weeds. And if it determines that the lettuce isn’t growing optimally, it will spray it too.

Although machine learning benefits machine-learning, there are some challenge and limitations. “We should emphasize that machine learning does not result in the automation of the farmer’s job” (Cantor and Seibel, 2016). Machine learning is more of a complement to farmers rather than a replacement. Technology investment can also be a big investment for some farmers who prefer doing everything manually. Also, machine learning is relatively new in agriculture so it’s not 100% accurate. The hardest part is training the algorithm that puts the data to use. It takes time and requires more data to be collected.

Machine learning in agriculture is a relatively new technology, but with the recent advancements in technology and satellites, it will be very useful for farmers in the future. It will help improve crop yields and minimize herbicides used. Any challenges and limitations will be solved in the future with the steady advancement of technology.

Sources:
Cantor, M. & Seibil, M., (2016) How Machine Learning and Big Data are Driving Progress in Agriculture. Retrieved January 15, 2017, from

How Machine Learning and Big Data Are Driving Progress in Indoor Agriculture


Brokaw, A. (2016, August) This Startup Uses Machine Learning and Satellite Imagery to Predict Crop Yields. Retrieved January 15, 2017, from
http://www.theverge.com/2016/8/4/12369494/descartes-artificial-intelligence-crop-predictions-usda
Simon, M. (2016, May) The Future of Humanity’s Food Supply is in the Hands of AI. Retrieved January 15, 2017 from
https://www.wired.com/2016/05/future-humanitys-food-supply-hands-ai/
Rouse, M. (2016, February) machine learning. Retrieved January 15, 2017, from
http://whatis.techtarget.com/definition/machine-learning
Keywords: Agriculture, Machine Learning

Leave a Reply