A “Siri” assistant for the farm is knocking at the door: How artificial intelligence is about to take farming by storm
Blogpost by Nicos Keable-Vézina, Director, Precision Agriculture
The day when every piece of equipment and sensor will be connected to a farm management platform is within sight. Thanks to artificial intelligence and all data collected by their farm’s sensors and digital systems, producers will simply have to turn to their virtual assistant to know how their operation is doing, be warned of critical events coming before they take place and get recommendations to help them make optimized decisions. Although the scenario sounds like fiction, it is virtually knocking at the door. Voice assistants like Siri and Alexa have already entered our homes. However, transposing this reality to agriculture involves working with life. And as such, a host of historical, current and future data need to be taken into account, which makes it all the more challenging.
Democratizing artificial intelligence is the missing link to building the easy-to-use prescriptive tools and autonomous equipment we’ve been hearing about for over thirty years. When combined to other technologies such as web and mobile applications, farmers and other agriculture professionals can suddenly turn to these real-time, turnkey solutions without having to care about the intricate technology behind them. Applications are vast and will make their way in every aspect of farming. Here are some examples:
Improving soil thanks to artificial intelligence
Bacteria and fungi are essential for soil health, with some having a major impact on nutrient availability. Companies such as Indigo Ag are working on sequencing beneficial bacteria to inoculate seeds. The benefits are huge: increased crop resistance to drought, increased yields, and resistance to certain diseases. This all was made possible by applying machine learning to genomics, whereby the agronomic potential of each bacterium was assessed.
Robots as fruit-picking partners
Robot pickers have entered the market in response to the lack of manpower. These autonomous robots are based on several technologies, which enable them to find the right position, recognize the right fruits and pick them. Several of these robots, such as Harvest Croo, have reached the marketable stage with the promise of replacing the equivalent of 30 strawberry pickers.
Prescription algorithms and interconnected data
The API interconnection between databases facilitated the development of artificial intelligence algorithms capable of taking into account very local weather data, various agronomic parameters, expected grain prices and farm input, as well as plot variability as detected by imagery. All of this data can be analyzed in real time in order to deliver an optimal prescription for each season and each field plot, whether it is for prescribing nitrogen as our FieldApex tool does, or simply for locating problem areas with the help of a small mobile application developed by create4d which leverages augmented reality via a mobile phone.
Crop management and pest identification
Big players in the industry such as Granular, Climate and Xarvio have launched or acquired easy-to-use platforms in precision farming. Initially designed as farm management software, these platforms are beginning to offer pest identification tools with no other requirements than taking pictures with a mobile device. Pictures are then compared to image banks. This is a huge market, and independent players like Plantix also offer especially interesting solutions from this perspective. And that’s only a start. The huge amount of data collected by these tools will lead not only to the quick identification of pests and crop diseases, but also to more optimal pesticide prescriptions. Another novelty worth mentioning and expected to accelerate in the coming years is to integrate satellite imagery with these platforms to enable the identification of problem areas and then use a drone, a robot or a cell phone to accurately identify pests while minimizing travel.
Beating grain markets from home
To sell crops, several farm producers follow market prices and crop yield forecasts reported by the US Department of Agriculture (USDA). New platforms now help farmers to take positions on the markets without having to spend as much time as a normal trader would. Thanks to the assessment of yields made possible by satellite imagery, weather forecasts and algorithms, several platforms have started to offer turnkey solutions that give reliable information for producers to take positions on the market and sell their harvests. Run a quick search on the web to visualize existing offers.
Weed control is a time-consuming task for farmers and account for significant production costs. By using computer vision for weed identification, companies such as Blue River Technology now offer selective pesticide applicators that can spray weeds individually. Another company even pushed this concept to develop Ecobotix, an autonomous, solar powered robot capable of spotting and spraying weeds one by one.
Optimal management of herds and buildings
Livestock is not to be left behind. We’re speaking of controlled environments here, and this makes it possible to quickly collect the equivalent of hundreds of years of monitoring by collecting data from several herds at a time. As for henhouses, companies such as Intelia make them smarter by collecting data on weight, temperature and feed consumption. This enables farmers to obtain growth projections and take remedial actions along the way. With the help of the electronic scale that they have developed, farmers can predict the weight of chickens over a period of 14 days.
For their part, dairy farms also stand out with the entry of the virtual assistant Ida for producers whose cows carry a sensor around their neck. The device transmits the movements of the animal to a program, by which they are interpreted through artificial intelligence. By comparing the data collected and the actual behavior of the animal, the system learns to recognize various behaviors, namely chewing, resting on the ground, walking, eating and drinking. Behavioral monitoring has led to the development of algorithms that can detect the occurrence of mastitis and limping before these conditions become critical. The assistant is also able to identify ovulation peaks with a 90% reliability rate, as well as certain eating disorders. More is to be expected. Other initiatives such as the Virtual Farm Brain developed at University of Wisconsin promise to push these concepts much further.
How to become a leader of tomorrow
If you are an agronomist or farm producer, you’ve got to start getting familiar with these tools because they will quickly become essential to your daily life. In addition to technology companies and agribusiness, it is of utmost importance that farm producers discuss these new opportunities in their cooperatives and businesses. More than ever before, given the scarcity of expertise and the costs involved in developing new technologies, local players will need to work collaboratively and open up their databases to scientists and technology companies in order to contribute to the research and development efforts. For them to lose neither competitiveness nor independence in the face of the farming industry and international markets, farm producers will certainly have to insist on transparency in the use of data, without this issue becoming an obstacle to technology development.
If you are a developer of technology solutions, remember to include farmers, field agronomists from diverse backgrounds and entrepreneurs right from the beginning of your research projects. This will facilitate the adoption of your technology. The farm community holds a myriad of opportunities. However, there are strong barriers to entry. As the findings of a survey conducted by Croplife magazine in 2018 point out, two factors are key to the success of an agricultural innovation:
- innovations must be ready for immediate integration into current farming practices;
- the farm business adviser must be at the core of the integration effort with the farm producer, to leverage the existing bonds of trust between them.
Major technological advances bring their share of questions and uncertainties. It’s up to innovators to do it the right way. No matter how promising, disruptive innovations for agriculture must remain on a human scale and first and foremost pursue the goal to serve users. In this sense, what matters most is that the farm producer and the team of professionals who supports him be placed at the core of the development of artificial intelligence-based tools meant to serve their needs.
This blog post was an overview of the major innovations that reached the marketable stage. If you are an innovator with a developing business who applies artificial intelligence to agriculture, please contact me to introduce your business. We could write a future blog post together about the innovators of tomorrow. And who knows if this will be the spark of a new partnership with FieldApex or the company propelling the spin-off, Effigis Geo-Solutions.