AI in Agriculture

Oct 1, 2018 · 4 min read

Agriculture, as we all know, feeds the world — and the population of the world is growing. In 2017, it was 7.5 billion. By 2050 it could be 10 billion. The environmental strain that is being put on the planet by our growing population and industries, including agriculture, is leading to runaway global warming. Plastic waste is choking the oceans, chemical runoff is contributing to dead zones and threatening sea life. It is clear that something has to change in order to manage continued growth in a sustainable fashion. believes strongly that the application of artificial intelligence to agriculture could be one of the most important potential answers we need to increase crop yields and make livestock farming more efficient while reducing the total amount of chemicals, fossil fuels and waste produced by the industry.

Globally, agriculture is a $3 trillion industry that employs over 1.5 billion people, which is 20% of the world’s population. Today, family-owned farms produce 80% of the world’s food. In the past 100 years, crop yields in the developed world increased by 600% on average, allowing the food supply to rise dramatically with the growing population while keeping the amount of land under cultivation stable. At the same time, the percentage of people employed in the industry dropped precipitously — leading to further economic growth as large amounts of human capital were unlocked and freed to work in more productive industries.

These changes were made possible by numerous technological and scientific advances. They include the mechanization of agriculture, the invention and continual improvement of motorized combine harvesters and other agricultural machinery, the discovery of chemical fertilizers, advanced crop management techniques and continually improving seed genetics.

While these advances have revolutionized agriculture in the United States and Europe, much of the developing world still has a long way to go to catch up. Global corn yields, for example, are still only 52% of corn yields in the United States.

But as global agricultural yields increase, so too will the environmental challenges presented by the agricultural industry. Fresh water sources are fast becoming depleted, available agricultural land is becoming more scarce and soil quality is being degraded. According to the FAO (Food and Agriculture Organization of the United Nations), climate change will only exacerbate these issues. The population is projected to increase by 2 billion by 2050. By this time, however, only an additional 4% of arable land will come under cultivation.

In order to maintain our food security as well as environmental safety, it will be necessary to further increase crop productivity, decrease the use of water, fertilizer and pesticides. Agriculture will need to be more efficient in terms of use of inputs, including dangerous chemicals,

The promise of artificial intelligence in agriculture will be enabled by numerous other technological advances, including big data analytics, the internet of things, the availability of cheap sensors and cameras, drone technology and even wide-scale internet coverage on geographically dispersed fields.

By analyzing disparate data sources such as temperature, weather, soil analysis, moisture, and historic crop performance, AI systems will be able to provide predictive insights into which crops to plant in a given year and when the optimal dates to sow and harvest are in a specific area, thus improving crop yields without requiring the use of potentially dangerous and expensive additional fertilizers.

AI-powered predictive analytics could also warn of increased risks for pest attacks, allowing for the more strategic use of pesticides, reducing their overall requirements.

Machine vision technologies, a particular focus area for , have potentially revolutionary applications in agriculture. The use of machine vision could allow for new robotic harvesters which can ‘see’ and handle delicate fruits and berries.

Another application of machine vision currently being pursued by John Deere is in the precise application of weed killing chemicals. They have developed an autonomous robot which is able to distinguish crop plants from weeds using a deep learning computer vision model. The robot is able to administer precise amounts of weed killer only where needed, thus dramatically reducing the amount of these chemicals required to protect a crop.

Precision agriculture is another emerging field, which has seen numerous applications for artificial intelligence. It is based on the concept that variations exist in the environmental conditions on a more granular level than farmers were previously able to track. Using new technologies such as satellite imaging, distributed sensor data and autonomous drone surveillance of fields, researchers are able to train AI systems to optimize the use of inputs on a precise level — i.e. applying just the right amount of fertilizer, only in the part of the field that requires it, thus maximizing yields while preserving resources.

In livestock farming, is currently conducting field research and development on computer vision systems for identification and tracking of individual animals and even analysis of animal weight via an intelligent camera application, allowing for market weight optimization and improved profitability and time to market.

The promises of artificial intelligence in agriculture are numerous and of great importance to the world. Higher crop productivity and decreased use of water, fertilizer, and pesticides via application of artificial intelligence technologies can reduce the impact on natural ecosystems, and increase worker safety, which in turn keeps food prices down and ensure our food production system will keep pace with population while keeping our planet safe.

For more information on how to develop artificial intelligence systems for the agricultural industry, please contact .

By Angus Roven,

Neuromation Investor Relations Analyst


Distributed Synthetic Data Platform for Deep Learning Applications


Written by


Distributed Synthetic Data Platform for Deep Learning Applications