AI is revolutionizing every stage of the agricultural process from planning and planting to crop estimation and harvesting. It might not be a glamorous industry but it has long been at the forefront of technological advancement and the advancements of AI in farming are truly impressive. Artificial intelligence in agriculture is a far cry from generative AI like ChatGPT but it incorporates all sorts of technology from computer vision and machine learning to lasers and robotics.
For many, agriculture is an abstract concept. With more and more people living in cities, farming and the production of our food has become quite removed from our daily lives. But agriculture underpins almost everything we do, from the food we eat to the water we drink and it is intimately linked to current affairs like climate change, deforestation and wildfires.
Great work is underway to utilize AI across a wide spectrum of agricultural activities. Let’s take a look at some recent developments in the key stages of the farming process.
Using AI To Plant Crops and Trees
Problem: the Earth loses 26 million hectares of trees each year, with over 30% attributed to wildfires.
Solution: AI-powered seed planting drones can help undo this damage by planting trees in previously inaccesible lands.
Getting seeds in the ground is where it all begins. Each year farmers and other professional growers have to plant their new seeds. Although there is a science as to where and when you plant, it remains a hugely manual and time-consuming process. Farming and reforestation are undergoing wholesale improvement due to artificial intelligence.
Pioneering firms like Flash Forrest and AirSeed have developed incredible futuristic autonomous flying machines to speed up this process. They use drones, computer vision, GIS and mapping technology to fly over inaccessible and hazardous lands and identify appropriate planting locations. Then, using robotics, they drop seed pods to the earth. These drones can plant seeds 25 times faster than manual human planting and are revolutionizing how we approach the planting process.
The drones have a camera attached to their base. The machine then flies over the land and scans the ground for appropriate planting locations using computer vision (the same software behind facial recognition). Advanced machine learning algorithms are used to make these determinations. GIS and mapping software then record the desired planting locations, and either the same drone or a partner machine will drop or some cases fire the seed to the ground using advanced, highly accurate robotics. This explainer video from The World Wildlife Federation explains the process.
The company AirSeed has a lofty goal of planting 100 million seeds next year. Their focus is on trees rather than crops but the technology is the same. Through the use of computer vision-enabled drones, they can plant as many as 20,000 seed pods every day, to help replenish forests that were destroyed by fires.
Crop Fertilization Using AI
Problem: waterways in many parts of the world have become contaminated by excessive nitrogen levels and 12% of the world's arrible land is no longer usable.
Solution: Artificial Intelligence powered soil sensors allow farmers fertilize far more accurately, boosting yields, reducing pollution and increasing profits.
Next up in the growing process is the act of fertilizing, essentially feeding the young plants. Every plant requires the right mix of nutrients and minerals to grow, and some are more sensitive than others. One of the key nutrients is nitrogen, which is produced from ammonia by naturally occurring biomes in the soil. If the nitrogen levels are too low crop yields can be harmed. If nitrogen levels are too high then it can become a pollutant, especially when it runs off into water supplies.
Through the use of advanced sensors, known as chemPEGS, or “chemically functionalized paper-based electrical gas sensor,” machine learning algorithms can accurately predict soil nitrogen levels over the next week or so (generally up to 12 days). This allows farmers to fertilize more accurately and sparingly, reducing their costs and minimizing the risk of pollution and run-off.
Weeding and Pest Reduction Using AI
Problem: the earth loses 40% of annual agricultural crops to pests.
Solution: AI-powered technology can remove 200,000 weeds per machine per hour and reduce pesticide useage by 90%.
Weeding is the bane of any farmer’s existence and one of the primary applications of artificial intelligence in agriculture. The suite of technology here is quite similar to what we saw during the planting process, with computer vision and machine learning underpinning the automation.
However, instead of connecting this equipment to futuristic drones, these AI-powered weeding rigs are connected to normal-looking farm equipment and tugged along behind a tractor. Of course, beneath the surface, this technology is still cutting-edge.
One of the pioneers in this space is Blue River Technologies (read our in-depth case study here) with the groundbreaking Sea & Spray product. In simplistic terms, this technology uses facial recognition software to identify weeds and spray them with pesticides. This selective application of chemicals allows for a 90% reduction in pesticide use compared to a more traditional blanket spray approach, saving farmers money and reducing the amount of toxic pesticides being used.
Another approach is the laser-zapping technology of Carbon Robotics. The setup is very similar to See & Spray, but instead, the weeds are zapped out of existence by laser. This technology can remove up to 200,000 weeds per hour, something that would take 70 people to achieve using non-AI methods.
Harvesting Crops Using AI
The final part of the farming process is typically the harvesting of the crop. Some types of produce are easily harvested by machine, and generally the more delicate the product the harder it is to automate its harvesting. Combine harvesters have existed for decades, automating the relatively straightforward harvesting of cereals. Strawberries, kiwis and other such fruits were long considered beyond our technological capabilities as the fruits are so delicate. These are typically still picked by hand.
AI is changing all that. Once again computer vision is a key component of this process. A camera is used to scan the plants and identify pickable fruit. GIS is used to map out where this is. Robotics is used to gently grab or hoover up the fruit. This is hard to imagine but the below video from AbundantRobots illustrates the process.
Crop Yield Estimation With AI
Apart from automating the mechanical aspects of farming, AI is being used to improve the science of agriculture too. Understanding crop yields is a vital component of farming. It is also part of governmental and economic forecasting. It impacts food prices, can help predict and prevent famines and plays a big role in global financial markets.
Companies like Gro Intelligence use artificial intelligence to make highly accurate predictions about agricultural output. They utilize over 170,000 data sets in their models, everything from satellite imagery and soil samples to price data and the weather. Their proprietary models are then used to predict yield estimates from crops like Brazilian soy and US corn, measure the amount of arable land in West Africa, or predict if unseasonably high rains will damage European crops or cause urban flooding.
This does not change the processes of how we farm, but it does allow farmers, corporations and governments to make better decisions in terms of agricultural planning. This can help tackle some of humanities greatest problems from hunger and famine to flooding and wildfires.
AI Advancements In Farming Summary
Artificial intelligence is driving some real change in farming, and farming is helping advance AI. It is helping us make better decisions, plant more seeds, access more land, roll back deforestation, use less chemicals and make farming more profitable. Farming has long been at the forefront of technological advancements and in this era of artificial intelligence, agriculture is once again driving human development.
Σχόλια