Products with intelligence: how neural networks are being introduced into the food industry

Such data were cited at the end of September by Dmitry Chernyshenko, Deputy Prime Minister of the Russian Government, in his speech at a strategic session on the development of AI until 2030. The authorities are confident that artificial intelligence has colossal potential .

"The updated National Strategy for the Development of Artificial Intelligence until 2030 will help to reveal it. It includes measures to develop infrastructure, science and personnel, as well as support for implementation," said Dmitry Chernyshenko. Russian President Vladimir Putin

approved the National Strategy for the Development of AI until 2030 in 2019. Its main goal is to ensure the accelerated implementation of neural networks in the country's economy. Since last year, the government has been working on updating this strategy. The document provides for the implementation of a set of measures aimed at the widespread introduction of such technologies in economic sectors, the social sphere and the public administration system. When implementing the strategy by 2030, 5 thousand organizations will receive preferential access to computing power for AI development, 95% of all companies will use artificial intelligence , including thanks to government support. In addition, 300 research and applied projects will be launched, including with international participation, Chernyshenko said. According to a study by the Institute for Statistical Studies and Economics of Knowledge (ISSEK) of the National Research University Higher School of Economics, today two-thirds (65%) of organizations - users of AI-based solutions use such technologies in test (experimental) mode, studying and evaluating the capabilities of new business solutions. The most popular AI technologies are computer vision , speech recognition, recommender systems, biometrics and natural language processing. Most of all, neural networks are needed by companies to improve or create products and business processes. At the same time, organizations prefer mainly Russian products. The food industry was included in the number of priority sub-sectors for the implementation of artificial intelligence in 2023. Today, AI-based technologies are successfully used here at all stages - from production planning to delivery of finished products to stores.  Artificial intelligence helps to forecast demand, control compliance with sanitary standards, track the quality of raw materials, and monitor production. We tell you about some AI solutions in the practice of Russian food producers and what problems they solve. Sorting and quality control of raw materials







Raw materials are often heterogeneous, the conveyor may contain unripe or spoiled fruits, foreign objects, etc. When sorting manually, there is a possibility that a defective product will be sent for further storage. Artificial intelligence comes to the rescue. For example, the Russian company Mallenom Systems has developed the intelligent system VISKONT.Beet, which helps to assess the quality of SUGAR beets at the acceptance stage.

Using neural networks, the system evaluates the contents of the truck body for dirt, tops, chips and snow. Based on the data received, the raw materials are classified: processing or storage.

The cost of an error in accepting raw materials can be high, since sending low-quality beets for long-term storage can lead to the spoilage of a significant volume of raw materials and large economic losses for the enterprise, as explained on the developer's website. The accuracy of decisions made by an employee at the acceptance point is on average 60-70%, while the AI-based system does this with an accuracy of over 90%. The technology is used, in particular, at the Znamensky sugar plant of the Rusagro company in the Tambov region.

Monitoring the work

of workers on the conveyor line The Morigan.Lean video analytics system from the Russian company Mattler is designed for enterprises that use manual labor: packaging, packing, displaying, assembly. The solution based on computer vision monitors employee productivity, identifies errors or opportunities for labor optimization. Cameras are installed above the workplaces at the enterprise, the neural network sees every operation performed by the employee. The system allows you to receive statistics on the individual productivity of each employee online , monitors compliance with production process regulations, sanitary standards, keeps track of equipment downtime, etc.

The solution is used, for example, at the Mosselprom (Cherkizovo Group of Companies), Pavlovskaya (Russkoye Pole Management Company), and Yaroslavsky Broiler poultry farms. After implementing the "smart" system, enterprises note an increase in labor productivity by 15-50%.

"Only a few video cameras are installed in the work area, and the main units - industrial computers - are removed from the work area. The system can be easily transferred to any workplace. It is possible to manage and control processes at almost all main technological operations," says Sergey Tereshchenko, Deputy General DIRECTOR of Yaroslavsky Broiler.

Control over compliance with sanitary standards in production

Enterprises of the Damate Group of Companies for turkey processing in the Penza Regionhave been using neural networks to monitor compliance with sanitary standards for several years now. We are talking about the artificial intelligence system "Direktiva: Sanitation", which works in tandem with a mechanical hand wash. The AI ​​solution, which reduces the risk of infections and the transfer of bacteria to production, was developed by the Russian company "Connectom".

The system uses computer vision technologies. The camera recognizes the employee's face, and instructions on how to properly sanitize hands are displayed on the screens, in accordance with the European standard EN-1500. The amount of soap foam is recorded, the thoroughness and duration of movements are monitored. Depending on how well the employee's actions corresponded to the regulations, a status is set. The system will not allow the employee to pass through the turnstile into a clean production area if the hand sanitization was not performed according to the regulations.

"Damate" became the first Russian company to completely transfer control over the sanitization of employees' hands to artificial intelligence, thus reducing the influence of the human factor to a minimum.

At Damate, the system has been implemented and successfully used in daily operations for many years, says Olga Chebunina, Advisor to the General Director for Information Technology at the Damate Group of Companies: “For any food enterprise, sanitary safety, in particular hand cleanliness, is an extremely important point. When entering the production facility, employees are required to treat their hands according to a certain strict regulation. This is monitored by interactive screens, cameras and servers. A person can pass the turnstile blocking the way to the clean zone only after further drying and disinfecting their hands. Artificial intelligence works 24 hours a day, it is very difficult to deceive the system, so the hands of the staff are always guaranteed to be clean, and the products manufactured  are high-quality and safe.”