Grabbing point acquisition method of robot stable grasping object based on monocular vision

Abstract

The invention belongs to the field of robot control, in particular to a method for acquiring grasping points of a robot steadily grasping an object based on monocular vision, aiming at solving the problem of low grasping success rate of a robot to an unknown object in an unstructured environment. The invention comprises the following steps: obtaining the color image of the object and the environment based on the monocular vision sensor, and extracting the plane edge contour of the object; Based on the planar edge contours of four-fingered parallel manipulator and object, constructing the environment constraint domain of robot grasping system, and obtaining the lowest point of environment attraction domain, and then a plurality of candidate grasping points. Inputting the candidate grabbingpoints of each group into a grabbing point quality evaluation network to obtain the grabbing point quality of each group of candidate grabbing points; Selecting the output of the grab point corresponding to the maximum quality of the grab point. The invention improves the rapidity, accuracy and reliability of the grasping point identification, and improves the grasping success rate of the robot tothe unknown object in the unstructured environment.

Description

technical field [0001] The invention belongs to the field of robot control, and in particular relates to a monocular vision-based method for obtaining a grasping point for a robot to stably grasp an object. Background technique [0002] The automatic identification and stable and fast grasping of objects by robots plays a very important role in the realization of industrial production automation and the wide application of robots in industry, and is the prerequisite for the completion of automatic assembly of robots. At present, according to the nature of the object to be grasped, the robot grasping system is mainly divided into two directions, one is the grasping system based on the object model, and the other is the grasping system with the unknown object model. The grasping system based on the object model needs to obtain the geometric model of the object to be grasped in advance, and the grasping accuracy is high, but the operating environment is very structured and the ...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the above-mentioned problems in the prior art, that is, in order to solve the problem that the robot has a low success rate of grasping unknown objects in an unstructured environment, one aspect of the present invention proposes a robot stabilization system based on monocular vision. The grabbing point acquisition method for grabbing objects includes:

Images