Algorithms for objects precise localization based on visual sensors of industrial robots

The article deals with the problem of localization of objects in the image, which occurs when using industrial robots. A method is proposed that makes it possible to identify the object and determine its position in space, regardless of the angle of the survey. The developed algorithm includes the Hough transformation to determine the potential vertices of the object, filtering out false candidates based on clustering methods, and optimized enumeration of variants of model and object image matching. As a result, a prototype of the system was created, which allows (on an example of a cube) to determine the angles of rotation of an object in space based on its image.