An increasing number of tasks in agriculture are being delegated to smart systems. With the help of AI in Agriculture, agricultural producers can improve economic efficiency by reducing production costs and increasing yields.

Today, when making decisions, an agricultural producer has previously inaccessible sources of information: images of satellites and UAVs, readings of humidity sensors, ground weather stations, etc. At the same time, new monitoring and control systems are constantly appearing on the market, which offer individual, more accurate analysis and forecasting.

In 2014, 190,000 measurements were taken on smart farms to inform farmers. By 2050, the number of measurements will grow to 4.1 million per day. It is almost impossible to navigate this flow of information on your own.

One of the tasks of using artificial intelligence in the agrosphere is the generalization, analysis and processing of data from various monitoring tools, and the issuance of recommendations based on them.

Machine learning for monitoring fields

The Israeli startup Taranis provides accurate information about the condition of plants, allows timely identification of negative factors and gives advice on their prompt elimination. For monitoring, readings of field observation sensors, meteorological data, and aerial photography are used. The analysis uses ultra-high resolution images (up to 8 cm per pixel) from Mavrx.

Based on the analysis of the data, areas of crops with inhibited growth are identified, plant diseases, problems with pests are identified, the supply of plants with nutrients, potential yield, etc. …

Platform Watson Decision Platform for Agriculture from IBM advises farmers using remote sensing data. The user receives information about the presence of damage to corn as a result of diseases or pest attacks. The system analyzes the likelihood of such injuries based on hyperlocal weather forecasts and individual crop data.

IBM’s Watson will determine the type, quantity, and optimal time frame for the farmer to treat the affected area with pesticides. It will help in carrying out preventive treatment. Using the high-resolution plant activity index (HD- NDVI ), it will assess the condition of the plant, determine the necessary preventive measures (application of fertilizers, nutrients, etc.).

By combining moisture data (HD-SM) with terrain data and meteorological measurements, soil moisture dynamics are simulated. The farmer also receives a yield forecast, the dynamics of yield changes based on photographs and information from previous seasons, etc.

Farmers Edge’s Health Change Maps and Notifications artificial intelligence platform  informs the farmer about the efficiency of the equipment, the condition of the plants, the appearance of pests or diseases, nutritional deficiencies, etc. The program processes satellite images and sends messages to the user about possible risks and necessary measures.

Application Field Manager from Bayer gives recommendations based on satellite images and processing downloaded data. With the help of a mobile phone, the farmer can at any time receive information about the state of vegetation, the degree of protection, and the local weather forecast. The user receives messages about possible risks and recommendations on how to prevent them. 

The Hummingbird Technologies platform provides the farmer with information about the current state of crops. For the analysis, data from satellites, UAV images, information from ground-based monitoring facilities are used. Proprietary algorithms are used to interpret the data.

The company provides clients with information on the amount and condition of plant mass, the presence of weeds, the supply of plants with nitrogen, etc. On the basis of this information, maps of the differentiated application of fertilizers, nitrogen and plant protection products are developed, calculations are carried out to simulate the optimal irrigation regime. Using this information, an agricultural producer can increase yields by reducing the cost of production.

The advantages of satellite monitoring for efficient farming, as well as real cases of using this technology will be described in the “Precision Farming Guide”.

Smart spraying technologies

The active use of pesticides and agrochemicals in the agricultural sector has been going on for over 60 years. This practice is generally accepted and the legislation of most countries is loyal to the widespread use of such funds. However, the side effects of their use are obvious to everyone today.

A decrease in the amount of pesticides used, in addition to the economic effect, will have a beneficial effect on the state of the land fund. That will subsequently have a positive effect on the yield.

Autonomous system WeedSeeker company Trimble produces spot spraying of weeds. The system identifies weeds using LEDs that scan the surface in the red and infrared range.

The reflected light is automatically analyzed, when a plant is detected, a signal is sent to a nozzle, which is triggered exactly above it. The response time depends on the travel speed. Salvo injection makes it possible to efficiently operate the system in strong winds. In areas where weeds occur periodically, it is possible to save up to 80% of the active substance.

The same principle applies to the WEEDit precision spraying system , which identifies and treats weeds at a tractor speed of up to 25 km / h. Both systems offer the possibility of differentiated herbicide application with individual nozzles and sensors.

For differentiated fertilization, Trimble offers the GreenSeeker RT 200 system , which works in a similar way to the WeedSeeker. GreenSeeker is a touch-sensitive yield sensor that determines in real time the required amount of application. The system can also be used when applying nitrogen fertilizers, determining the availability of nitrogen in the soil.

Other developments include systems that automatically adjust the height of the spray boom and enable and disable nozzles. Technology HawkEye company Raven  via the valve system provides a uniform pressure, giving the required amount of active ingredient for each nozzle.

Raven hawkeye

Smart Weed Control Projects

Scientists continue to work to improve existing methods of weed and pest control. Smart Garden Sprayer Tested in India . The sprayer installed on the tractor, using a system of ultrasonic sensors, determines the size of the tree and the distance to it. The information obtained is analyzed and affects the power of the jet and the amount of sprayed substance. Testing has shown high efficiency of the system, while reducing the consumption of plant protection products to 26%. 

Bayer and Bosh are developing Smart Spraying technology . It will differ from commercially available systems due to its ability to distinguish weeds from crops. It is assumed that the system will “recognize” the weed and determine the type and required amount of pesticide, taking into account the programmed application parameters.

” Killer Weed ” from the company EcoRobotix able to independently move around the field, dotted by identifying and treating weeds found. The developers claim that the technology will reduce the amount of herbicide use by 20 times.

Rapid identification of plant diseases

Self-evolving AI helps farmers identify diseases, determine treatment, and assess potential damage. Due to multiple calls, the library of such services is constantly expanding, the number of detected diseases is growing. The farmer only needs to photograph the affected part of the plant and in a few seconds he will have complete information about the cause of its occurrence.

Peat’s Plantix mobile app gives users the ability to diagnose over 60 plant diseases.

The application contains a huge library of images, which are sorted by types of plants, bacteria, diseases, etc. The algorithm for identifying diseases is improving with an increase in the number of downloaded images.

The Xarvio digital platform, developed by Bayer and acquired by BASF last year, offers users the Scouting app .

Scouting helps to diagnose diseases, injuries, disorders of plant development based on processing photographs. Currently, the application recognizes 17 diseases. In the images of yellow traps, the algorithm identifies dangerous ones and analyzes their number. Weeds are identified with an accuracy of 32–99% (depending on the phase of plant growth), and the degree of nitrogen supply to plants is analyzed. The application sends notifications in case of detection of dangerous diseases or pests near the user’s areas.

Ukrainian startup Field Monitoring plans to compete with foreign applications. Its algorithm based on a mathematical model of neural clusters is in the long term more efficient than competitors. The function of processing UAV images with the detection of diseases and the determination of the coordinates of the location of the affected plants is planned.

Other smart systems

The startup Uptake processes the data of the sensors installed on the equipment in order to optimize work processes. Guided by the recommendations of the program, the farmer will be able to improve the efficiency of using the equipment available on the farm.

American scientists are developing a precision irrigation system that can fulfill the task of ensuring the optimal amount of moisture. The project is intended to solve the problem of controlling water consumption at the local level (for each plant). The RAPID system (robotic precision irrigation system), consisting of adjustable emitters installed on the drip irrigation system, will ensure maximum irrigation accuracy.

The trend towards organic farms is forcing the search for an alternative to the use of agrochemistry in weed control. In many countries, there is a shortage of labor resources. Plus, human labor makes production too expensive. That is why active work is underway to introduce artificial intelligence technologies into the agricultural sector.