The joint project of "machine vision" "RT-Tekhpriemki" and Videomatrix is included in the open library AI Russia
Customers: RT-Tekremenka
Contractors: VideoMatrix (Videomatrix) Product: Vmx Dequs: ISProject date: 2020/03 - 2020/12
|
2021: Inclusion of the draft AWS of the final inspection of steel sheets in the AI Russia library
The project RT-Tekhimki"" of the automated workplace of the final inspection of steel sheets (ARMIL), developed and implemented jointly the Russian with the company (Videomatrix resident), Skolkovo for metallurgical production fell into the open library - AI Russia the project. This was reported Alliance in Artificial Intelligence on April 9, 2021 in RT-Tekhpriemka.
As of April industries , 13 innovative solutions are presented in the library in the region. Among them, a project was added to create a complex based on artificial intelligence, neural network technologies computer vision and to control the quality of steel supplied to enterprises of the aviation cluster of the State Corporation. Rostec
"The popularity of artificial intelligence is growing every year. We consider the demonstration of the practical application of technology to the business community one of the priorities of the Alliance. The goal of AI Russia is to present business cases with proven effectiveness for inspiration and further replication. The library will develop and be regularly replenished, and we are confident that very soon the project will become a "mirror" of the Russian AI market, "said Kirill Bulatov, General Director of the Alliance in the Field of Artificial Intelligence. |
"The recognition of our project by the Alliance's expert community in the field of artificial intelligence is highly appreciated by us. We plan not to stop there and introduce this software and hardware complex to ensure the quality of products of other industrial enterprises, as well as implement other projects, for example, the introduction of advanced machine learning at enterprises for accurate forecasts of equipment failure, "said Denis Kononchuk, First Deputy General Director of RT-Tekhpriemka JSC. |
2020: Completion of pilot operation of the automated workplace of the final inspection of steel sheets
JSC RT-Tekremenka"" (part of the State Corporation) Rostec together the Russian with the company LLC "" (VideoMatrix resident) Skolkovo completed the pilot operation of a prototype automated workplace for the final inspection of steel sheets for Chelyabinsk metallurgical production. This became known on October 27, 2020. The purpose of the tests carried out since the beginning of August 2020 was to determine the suitability of ARMIL, in other words Vmx of the system on the base and, machine vision for artificial intelligence pilot operation.
Technical acceptance specialists located at the Chelyabinsk production site, together with the developers of Videomatrix, trained the neural network for three months to improve the accuracy of defect detection. Based on the test results, it was established that the Automated Workstation for Inspection of Sheets (ARMIL) meets the requirements of the technical assignment of JSC RT-Tekhpriemka for the development, manufacture and supply of its prototype for metallurgical production.
The ARMIL complex based on artificial intelligence for steel quality control has proven its effectiveness and operability throughout the test period, so we are ready to move on to the next stage, namely to pilot operation of this plant and demonstration of its capabilities to the consumer. I note that as a result of tests and training, the system was able to recognize 97% of more than 20 types of defects, - said First Deputy General Director Denis Kononchuk. |
As of October 2020, the software and hardware complex controls the quality of corrosion-resistant VNS-9-Sh steel developed by the All-Russian Research Institute of Aviation Materials (VIAM). It is used to manufacture components of the carrier system of Russian helicopters.
The hardware and software complex combines artificial intelligence, neural network technologies and computer vision. For even more accurate operation of the system, additional training and an increase in the amount of statistical data in the data set are required. The technical acceptance specialists will continue to scan the VNS9-Sh sheets to further collect information on defects and self-study the system until the end of 2020.
The solution initiator, architecture designer and software implementation operator is JSC RT-Tekhpriemka. The solution is based on software for high-speed server-based industrial video analyticsVmx Dequs. ARMIL as a software and hardware complex is its own know-how and the development of Videomatrix LLC.