Specialists from the North Caucasus Federal University (NCFU) and St. Petersburg Mining University (SPGU) have trained a robot to recognize and remove defective parts from a production line. This was reported by the university's press service.
To achieve this, the scientists modernized a robot manipulator developed in St. Petersburg.
The most challenging aspect was achieving stable recognition rates for defective parts by the robot and ensuring that the time allocated for processing one object was constant. As a result of applying computer vision technology, it was possible to achieve high speed in performing all three operations: recognizing the part, identifying the defect, and removing the part.
Neural networks were used to train the robot. A special algorithm was developed using Python (programming language). It allowed for fine-tuning the interaction between the robot's mechanism and the computer vision system.
The goal of the developers is to minimize the number of defects in production and relieve employees of monotonous work.
Earlier, www1.ru reported that Moscow scientists created a multifunctional warehouse robot with integrated AI.
Read more on the topic:
With AI and technical vision: Rostec unveils tests of the Impuls-M combat robot for the first time
Replacing humans: Pixel cleaning robots will appear in Moscow parks
Russian robot dog from Moscow State University will go to schools and universities
Now on home
