Analysis, monitoring and forecasting: the main tasks of AI
AI in agriculture is used everywhere:
AI is used in precision farminga modern and technological approach to agricultural production, which takes into account heterogeneity within a single field. Typically, a field consists of sections with different soil characteristics (the ratio of clay, sand and silt, the amount of phosphorus, nitrogen, potassium, and so on). This must be taken into account when sowing, processing and harvesting.
Since agriculture is not a high-margin area, and many types of activities can be planned unprofitable, the ability to reduce costs helps the company survive.
"AI is used in precision farming - this is the name of a modern and technological approach to agricultural production"
Soil analysis
Usually, to find out what the soil is made of, you needtake samples in different areas. For an agricultural holding that manages 10-20 thousand hectares, this is expensive and labor-intensive. AI comes to the rescue - it analyzes the field as a first approximation using images from UAVs or satellites, determines the type of soil, the amount of humus in it and the ratio of different elements.
The main satellite imagery providers are the Sentinel family of satellites launched by the European Space Agency, the American Landsat program and Roscosmos.
Differentiated field processing
If the field is not uniform, it is divided intoseveral plots. Usually, the NDVI index is used for this, which is calculated from images from satellites and UAVs and allows you to find out what condition the plants are in in different parts of the field. Based on this index and other indicators, it is possible to develop maps for differential tillage of fields (plowing, fertilizers, treatment with plant protection products). This will save on fertilizers, fuels and lubricants and plant protection products. Differentiated irrigation, spraying of weeds and crops also works.
AI models tell the farmer or agronomistwhen you need to plant a crop and harvest it, apply fertilizer. Usually it looks like a reminder of the need to take some action, the decision is made by a person.
Forecasting the incidence and occurrence of pests
AI can predict the occurrence of diseases andpests in the field. As a base, either weather data is used (when the farm does not have special equipment), or information from sensors, cameras, and high-resolution drones. Such an analysis finds diseases at an early stage or recognizes pests before they spread across the field and saves the crop.
Harvest forecasting
AI helps businesses estimate future harvests.This information is needed for the entire season of agricultural work, not only to make plans, but also to correct them if something went wrong. Based on historical data, algorithms can be used to build a yield map: it will show how much the farm will collect from each part of the field, depending on its agrochemical and agrophysical indicators, relief. With the next landings, you can rely on this data.
“AI helps businesses assess future harvests”
Identification of problem areas
With AI you can identify problem areas:dry and flooded areas of the field. After the initial analysis, a farmer or agronomist can go to the site and conduct research using other instrumental methods. Finding problem areas is useful for insurance purposes.
Assessment of investment attractiveness
Sometimes, according to documents, the fields are designed asagricultural land, but in fact long overgrown with shrubs or trees. Sometimes the land is used incorrectly, and the soil layer is depleted - then soil reclamation is needed, and this is additional investment. Based on historical data and satellite images, you can determine the state of the field and roughly estimate how much you need to invest in order for it to start making a profit.
State monitoring and control
The area of unused agricultural land in the Russian Federation is almostin 44 million hectares. Often the land is listed as agricultural land, but in their place there are forests, buildings, landfills (the number of illegal landfills recorded at the end of 2021 in Russia exceeded 15 thousand, which is 30% more than it was at the end of 2019). Sometimes the state subsidizes and issues grants for the development of agriculture in a certain area, but the recipients do not use the land.
Control all processes personally by sendinginspectors for each field, on large volumes it becomes impossible. We need automation tools. AI allows you to find out what is happening on a particular piece of land and plan further actions.
“The area of unused agricultural land in the Russian Federation is almost 44 million hectares”
Specifics of agriculture in Russia
The Russian market has four features that affect the level of AI adoption.
Low digitalization.The Russian agro-industrial complex has fewpenetration of IT technologies. The Ministry of Agriculture of Russia supported the digitalization of the Russian agro-industrial complex five years ago, but only 5% of companies in the agricultural sector took advantage of this. And this is not AI, but simple technologies like reporting automation.
Large agricultural holdings have a higher level of digitalization.They more often use differentiated application of fertilizers, are interested in predicting yields, monitor the condition of fields and autonomously control equipment. But even they use scattered solutions that cover one or two needs.
There are few complex products on the Russian market,which combine all the information and create a digital twin of the agricultural holding, although the demand for this is high. Comprehensive monitoring of soil conditions increases the productivity of agricultural crops by at least 20%.
Fragmentation of these agricultural enterprises.Another feature of the Russian market - there is nocommon data formats and protocols for their transmission. Because of this, information for AI is stored in a fragmented form and is difficult to analyze. Sometimes important information is not available electronically at all.
There is little equipment of our own production.Before the imposition of sanctions in Russia, they usedWestern technology - for example, the American company John Deere. For its maintenance, imported spare parts are needed, and the firmware can be replaced only in official service salons. Due to the fact that the company stopped its activities in Russia, it will soon be impossible to use their machines.
There are domestic manufacturers in the country"hardware" and software for the agro-industrial complex. For example, the company Cognitive Pilot, which develops high-end "smart" control systems for agricultural machinery. But in order to scale these solutions and increase the number of domestically assembled equipment with installed Russian modules, time is needed.
Difficulties with retraining specialists.Many farmers and agronomists learn about the exactagriculture only after refresher courses. Specialists may store information about their fields on paper or in their head, rather than in an information system. When a person retires or moves to another agricultural enterprise, this knowledge has to be restored from scratch. At the same time, the average age of a farmer in the world is 55 years, this situation roughly corresponds to Russian reality (although it is declining). Some are ready to retrain, but many do not accept new or cannot afford further education.
Despite all the difficulties, the prospects forRussian agriculture is good, because Russia has huge land resources. The area of the land fund of the Russian Federation exceeded 1.7 billion hectares, of which about 22% is agricultural land. The question is that these lands need to be properly cultivated - and artificial intelligence will help in this.
“Many farmers and agronomists learn about precision farming only after advanced training courses”
AI instead of or together with a person
Experts have been debating for years whether AI canreplace a person. Of course, it will help medium and large companies make decisions and save money due to the processing of huge amounts of information, visualization of findings, quality recommendations and analytics. But so far there are no products on the market that trust AI for expert assessment, goal setting, planning and task control.
There are problems in the legal area:it is necessary to form a legislative framework for regulating activities using AI, to determine who will be responsible for its mistakes. And a person is not psychologically ready to give up control over technology.
But new professions will appear, and work will change andbecome more intelligent. By 2025, 97 million new jobs will be created because people, machines and algorithms will increasingly work together, predicts the World Economic Forum. With the development of technology, there will be a need for more qualified personnel with digital skills.
The agricultural industry will partially shiftto offices in the cities, from where they will manage what is happening in the fields. Routine operations are automated, but the person from the decision-making chain is not going anywhere.
Read more:
Plant on Mars produces oxygen at the rate of an average tree
Physicists have cooled atoms to record temperatures. They are a billion times colder than outer space.
A small dinosaur "turned" into a gem. He is almost 100 million years old