Method and system for visual servo controlling and equipment

Abstract

The invention discloses a method for visual servo controlling. The method comprises the steps that a training sample is obtained, a preset fuzzy neural network is trained by using the training sampleto obtain a visual servo controlling model; input expecting image information is received; the expecting image information is input to the visual servo controlling model, and current image informationof a target work piece is collected through an image collecting device; the difference of the current image information and the expecting image information is converted into a controlling signal of amanipulator, and the manipulator is controlled to execute the controlling signal to perform the next movement. According to the method, Jacobian matrix inversion is omitted to greatly improve the speed of visual servo controlling, complicated marking is not needed, the precision of visual servo controlling is effectively improved, and the motion controlling of a robot on automatically grabbing and transporting casting pieces is achieved. The invention further provides a visual servo controlling system, equipment and a computer storage medium which can achieve the above advantages.

Description

Technical field [0001] This application relates to the field of robot control, and in particular to a method, system, equipment and computer-readable storage medium for visual servo control. Background technique [0002] The research of robotics has a history of decades. In this process, the structure of the robot has been in the process of continuous development from simple to complex, from single to multiple, from low-level to high-level. With the development of technology, those technologies that restrict robots from adapting to complex environments are gradually overcome. In this process, the introduction of sensors makes it possible for robots to perceive complex environments. A variety of sensors provide a large amount of environmental information for the robot. In order to better imitate the mechanism by which humans obtain environmental information, people have introduced visual sensors for robots. Therefore, the control strategy between the robot vision sensor and th...

AI Technical Summary

Problems solved by technology

However, the extraction of these local geometric features is directly related to the shape and environment of the target object, and the algorithm for feature extraction and matching is also time-consuming.

Moreover, when the target object changes, different geometric features need to be used. Even if the pose of the same object relative to the camera coordinates has changed, the selection of features will also need to change accordingly, so the speed is very slow.

Secondly, the online calculation process of the real-time image Jacobian matrix is ​​cumbersome and time-consuming. Common dynamic estimation methods will cause large estimation errors, and error compensation is required.

Images