One of the most common computer vision tasks

[tanjimajuha20]

Scientists from the National Research University "MPEI" have developed a technology that allows accelerating the creation and implementation of computer vision systems for robotic systems, the aerospace industry and medicine.

is determining the position of spain whatsapp resource an object in a closed space. A striking example is determining the relative position of an object in a warehouse or the relative position of a patient and medical instruments and devices. This task is often solved using code markers - special objects that are easily recognized in an image. However, the accuracy with which the position of the markers is determined is difficult to estimate in advance. Usually, a working model is made and tested for this purpose.

The new development allows for the preliminary assessment of the required characteristics of computer vision system elements (primarily digital video cameras: resolution, matrix size, lens focal length, etc.) without the need to make working models or conduct expensive tests, which reduces the cost and time required to develop such systems.

"The technology developed by MPEI scientists is based on modern machine learning methods. The results of work in this area make it possible to reduce the high dependence on foreign technologies and accelerate the creation of domestic optical-electronic measuring systems," said Nikolai Rogalev, Rector of MPEI.

The technology, which can be used in the design of machine vision systems based on code markers, was developed by a research group from the V.A. Fabrikant Department of Physics at the National Research University "MPEI" as part of the implementation of the scientific research program "Priority 2030: Technologies of the Future" in 2022-2024.