Time for artificial intelligence. Belarusian scientists offer robots and devices based on neural networks to farmers
Topic news In recent years, the developments of the Scientific and Practical Center of the National Academy of Sciences of Belarus on agricultural mechanization are increasingly appearing at high-tech exhibitions and winning prestigious scientific competitions. Scientists offer farmers robots and devices based on neural networks. Help both yourself and your neighbor - There is a change in both approaches and technologies, states the General DIRECTOR of the Scientific and Production Center for Agricultural Mechanization, Candidate of Technical Sciences, Associate Professor Dmitry Komlach. – We are working to minimize the costs of technical re-equipment of agriculture, to speed up the time of operations and increase yields in all weather conditions. The development of various technologies makes it possible to obtain high productivity, minimize the impact on the soil, and maintain its fertility. The SPC has formed a plan for the development of machines for livestock breeding and crop cultivation until 2030. The document, approved by the Council of Ministers, outlines the domestic production of equipment for four Belarusian climatic zones. In addition, it is planned to create units for other countries. Thus, the SPC has already developed more than 65 types of machines for potato production: from planting equipment to lines for technological sorting and packaging of tubers, peeling and vacuum packaging. Developments are carried out jointly with other research and development centers and institutes that are part of the National Academy of Sciences; cooperation with Russian and Chinese colleagues is actively used, which allows products to be exported . “We work closely with the All-Russian Institute of Mechanization (VIM),” notes Dmitry Komlach. – Three programs of the Union State are at a high stage of readiness: “Digital Farm”, “Intelligent Gardening” and “Compound Feed-3”. All of them are related to high technology. The third program involves, jointly with the Belarusian National Biotechnology Corporation, the development of lines for the preparation of extruded feed, the digestibility of which will rise above 80%. Today it is at the level of 30-60%. “We pay great attention to training highly qualified personnel,” notes Dmitry Komlach. – Today we have 14 candidates and three doctors of science. Four candidate and one doctoral dissertations are being prepared for defense. Smart house for cows In 2021, SPC developed a hardware and software complex for the system of identification and monitoring of the physiological state of animals (IKFS), a domestic analogue of foreign systems. – We created the complex on the instructions of the Ministry of Agriculture and Food. Co-executor is the Scientific and Production Center of the National Academy of Sciences of Belarus for animal husbandry,” says Evgeniy Zhilich, HEAD of the laboratory for mechanization of MILK and beef production processes. – At the moment, 70 sets and about 15 thousand collars for cattle have been produced
. An import-substituting development has been obtained that allows one to abandon foreign transceivers and herd management systems.
IKFS collects and stores all information about the herd in a computer system. Including each cow's identification number, breed code, date of birth, current lactation number, disease codes and much more. In fact, a livestock specialist can study an electronic dossier on an animal at any time . Today IKFS is two to three times cheaper than its foreign analogues.
The Belarusian transceiver from a collar on a cow transmits information over a distance of up to 90 meters, is resistant to interference, and has a wireless charging function for the built-in battery. IKFS components are Belarusian and Russian, software is domestic. To create it, we conducted special research: we tracked the behavior of cows and rejected some of the indicators.
The main advantage of IKFS is that the system is the heart of a smart farm, to which add-ons can be added. So, in 2023, a device was developed that determines the pre-mastitis condition in cows. It consists of a thermal imager, sensors and software. Accuracy of work – 95%. The device will go into production in the near future. Thermal imaging matrices linked to IRFS at the entrance to the milking parlor will allow you to identify a sick animal and isolate it from the herd. The machine identifies the disease faster and better than a person, allows you to noticemastitis in the early stages, cure as soon as possible.
“With IRFS and thermometry, which allows us to determine the pre-mastitic state of animals, we entered the top 10 results of the activities of scientists of the National Academy of Sciences of Belarus for 2023 in the field of fundamental and applied research,” the scientist says proudly. – This year we are finishing the development of two more projects.
The first is the positioning system of the milking robot, or its “arm”, which is cheaper than foreign analogues. The robotic arm puts on and takes off the glasses itself and milks the cow, which eliminates the labor of the machine milking operator.
Work is also underway to create an intelligent microclimate system for cooling barns in the summer. At temperatures above 25 degrees, milk yield drops by 30-50%. The autonomous system, based on the weather forecast, will be able to cool the premises in advance. It will also control the level of carbon dioxide and light.
“In 2024, the development of an intelligent feed distribution complex will be completed,” Evgeniy Zhilich shares his plans. – A distributor with loaded components approaches each age and sex group of animals, checks the recipe in the IKFS and issues food that is ideal for them.
The laboratory also has an OM-1 feed pusher. Developed in the shortest possible time, it is at least five times cheaper than imported analogues. Each APPLE is “photographed” seven times . “When creating our laboratory, the task was set to mechanize the horticultural industry, which uses the largest share of manual labor,” says Anton Yurin, head of the laboratory for mechanization of the cultivation of fruits and vegetables, candidate of technical sciences, associate professor. – The first development was the creation of a platform that facilitates the work of harvesting apples and pruning fruit trees - a self-propelled universal unit ASU-6. Its use increased productivity during harvesting by two to three times, and when pruning – by five to six times. The transporting parts are completely mechanized; the picker just needs to carefully remove the apple from the tree and place it on the conveyor belt. Photo unit Another “apple” unit is the KUV-1.8 complex for harvesting fruit tree branches. When the same apple trees are pruned annually, a large number of branches are formed. Previously, five or six people collected them by hand, loaded them onto a trailer and hauled them away for disposal. It was necessary to burn 7–10 tons of branches per hectare, and this is a huge amount of organic matter that could serve as fertilizer. KUV-1.8 collects branches into a windrow, then crushes them into fiber, scattering chips over the surface.
For work in warehouses, a technological sorting and packaging line for apples LSP-4 has been developed. Previously, the process was done manually, with a daily norm of 400–600 kg per person. Each apple had to be assessed by size and looked for for defects. Sorting was a subjective process, which after an hour or two was aggravated by the worker’s fatigue.
LSP-4 does not get tired. It is equipped with a technical vision system, which is based on a video camera with backlight. The module takes photographs of fruits passing under it, rotating during transportation. The photos go to the control unit, where the computer compares them with the samples. During the time the apple moves along the tape, it is photographed seven times. This allows the neural network to identify defects - after the “photo shoot” the apple is assigned the appropriate class, and the flows are separated.
The belt at a speed of 0.3-0.4 m/s allows you to sort up to two tons of fruit per hour. A technical vision system with an artificial neural network operates under the control of a program written at the Joint Institute for Informatics Problems of the National Academy of Sciences. To write it, scientists created a training sample of defects - about 4 thousand photographs.
Artificial intelligence can distinguish defects from natural depressions on the surface of the fruit.
Among the laboratory's developments are the KPYA trailed semi-row berry harvester, the OD-2 double-row sprayer for berry fields, and the Zubr double-row boom sprayer. Currently, work is underway on a machine for cutting the crowns of berry bushes, primarily remontant raspberries and currants.
Alexey GORBUNOV, photo - provided by the Scientific and Production Center of the National Academy of Sciences of Belarus for Agricultural Mechanization
. An import-substituting development has been obtained that allows one to abandon foreign transceivers and herd management systems.
IKFS collects and stores all information about the herd in a computer system. Including each cow's identification number, breed code, date of birth, current lactation number, disease codes and much more. In fact, a livestock specialist can study an electronic dossier on an animal at any time . Today IKFS is two to three times cheaper than its foreign analogues.
The Belarusian transceiver from a collar on a cow transmits information over a distance of up to 90 meters, is resistant to interference, and has a wireless charging function for the built-in battery. IKFS components are Belarusian and Russian, software is domestic. To create it, we conducted special research: we tracked the behavior of cows and rejected some of the indicators.
The main advantage of IKFS is that the system is the heart of a smart farm, to which add-ons can be added. So, in 2023, a device was developed that determines the pre-mastitis condition in cows. It consists of a thermal imager, sensors and software. Accuracy of work – 95%. The device will go into production in the near future. Thermal imaging matrices linked to IRFS at the entrance to the milking parlor will allow you to identify a sick animal and isolate it from the herd. The machine identifies the disease faster and better than a person, allows you to noticemastitis in the early stages, cure as soon as possible.
“With IRFS and thermometry, which allows us to determine the pre-mastitic state of animals, we entered the top 10 results of the activities of scientists of the National Academy of Sciences of Belarus for 2023 in the field of fundamental and applied research,” the scientist says proudly. – This year we are finishing the development of two more projects.
The first is the positioning system of the milking robot, or its “arm”, which is cheaper than foreign analogues. The robotic arm puts on and takes off the glasses itself and milks the cow, which eliminates the labor of the machine milking operator.
Work is also underway to create an intelligent microclimate system for cooling barns in the summer. At temperatures above 25 degrees, milk yield drops by 30-50%. The autonomous system, based on the weather forecast, will be able to cool the premises in advance. It will also control the level of carbon dioxide and light.
“In 2024, the development of an intelligent feed distribution complex will be completed,” Evgeniy Zhilich shares his plans. – A distributor with loaded components approaches each age and sex group of animals, checks the recipe in the IKFS and issues food that is ideal for them.
The laboratory also has an OM-1 feed pusher. Developed in the shortest possible time, it is at least five times cheaper than imported analogues. Each APPLE is “photographed” seven times . “When creating our laboratory, the task was set to mechanize the horticultural industry, which uses the largest share of manual labor,” says Anton Yurin, head of the laboratory for mechanization of the cultivation of fruits and vegetables, candidate of technical sciences, associate professor. – The first development was the creation of a platform that facilitates the work of harvesting apples and pruning fruit trees - a self-propelled universal unit ASU-6. Its use increased productivity during harvesting by two to three times, and when pruning – by five to six times. The transporting parts are completely mechanized; the picker just needs to carefully remove the apple from the tree and place it on the conveyor belt. Photo unit Another “apple” unit is the KUV-1.8 complex for harvesting fruit tree branches. When the same apple trees are pruned annually, a large number of branches are formed. Previously, five or six people collected them by hand, loaded them onto a trailer and hauled them away for disposal. It was necessary to burn 7–10 tons of branches per hectare, and this is a huge amount of organic matter that could serve as fertilizer. KUV-1.8 collects branches into a windrow, then crushes them into fiber, scattering chips over the surface.
For work in warehouses, a technological sorting and packaging line for apples LSP-4 has been developed. Previously, the process was done manually, with a daily norm of 400–600 kg per person. Each apple had to be assessed by size and looked for for defects. Sorting was a subjective process, which after an hour or two was aggravated by the worker’s fatigue.
LSP-4 does not get tired. It is equipped with a technical vision system, which is based on a video camera with backlight. The module takes photographs of fruits passing under it, rotating during transportation. The photos go to the control unit, where the computer compares them with the samples. During the time the apple moves along the tape, it is photographed seven times. This allows the neural network to identify defects - after the “photo shoot” the apple is assigned the appropriate class, and the flows are separated.
The belt at a speed of 0.3-0.4 m/s allows you to sort up to two tons of fruit per hour. A technical vision system with an artificial neural network operates under the control of a program written at the Joint Institute for Informatics Problems of the National Academy of Sciences. To write it, scientists created a training sample of defects - about 4 thousand photographs.
Artificial intelligence can distinguish defects from natural depressions on the surface of the fruit.
Among the laboratory's developments are the KPYA trailed semi-row berry harvester, the OD-2 double-row sprayer for berry fields, and the Zubr double-row boom sprayer. Currently, work is underway on a machine for cutting the crowns of berry bushes, primarily remontant raspberries and currants.
Alexey GORBUNOV, photo - provided by the Scientific and Production Center of the National Academy of Sciences of Belarus for Agricultural Mechanization