An industrial robot vision system and teaching method thereof

Abstract

The present invention provides an industrial robot vision system and a teaching method thereof. The vision system includes a camera, a vision controller, an Ethernet switch, a robot controller, and a robot teaching device; wherein, there are multiple cameras, which are set in a working scene Any fixed position in or / placed at the end of the robot arm to move with the robot arm, the vision controller and the robot controller communicate through the Gigabit Ethernet network through the Ethernet switch, and the robot controller is connected to the robot teach pendant. The invention organically integrates visual collection, labeling, training and recognition systems, reduces system complexity and improves reliability. And through the method of teaching instead of programming, it is convenient to add visual recognition capabilities to industrial robots. This method does not require users to have professional machine vision domain knowledge. It reduces the difficulty of using machine vision products for industrial robots in the fields of object recognition, quality inspection, position detection, process inspection and safety protection.

Description

technical field [0001] The invention relates to an industrial robot vision system and a teaching method, belonging to the technical field of automation control. Background technique [0002] In the prior art, the application of visual recognition systems for industrial robots is becoming more and more common, but users need to perform complex, professional and long-term software programming in advance, requiring users to have professional knowledge in machine vision. Existing visual recognition systems generally use traditional algorithms such as feature matching or template matching to realize object recognition. This type of algorithm basically requires a solid-color background and is sensitive to illumination changes. The application scenarios are basically pure industrial occasions, which need to artificially limit the complexity of the scene and the change of illumination. Long deployment time and high cost. If the traditional feature matching method is used for iden...

AI Technical Summary

Problems solved by technology

[0004] In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a machine vision teaching method for industrial robots, which solves the problems of complex systems and the need for user programming.

Images