A method and system for estimating the pose of a target object, and a robot

Abstract

A pose estimation method and system of a target object, and a robot, wherein the pose estimation method includes: acquiring an image to be processed; inputting the image to be processed into a target detection network to obtain a target detection result image; inputting the target detection result image into training A good view reconstruction model is used to obtain a 3D reconstructed image. The 3D reconstructed image includes three channels, which are used to represent the 3D coordinates corresponding to the pixels; the transformation matrix is ​​calculated according to the 2D coordinates of the pixels and the corresponding 3D coordinates; The effective axis angle and the equivalent rotation axis can be used to obtain the pose of the target object. Because the view reconstruction model obtained through training is used to establish the mapping relationship between the two-dimensional coordinates of the pixels in the image and the corresponding three-dimensional coordinates, the three-dimensional reconstruction image covering the target object can be obtained through the view reconstruction model, and the three-dimensional coordinates corresponding to each pixel are obtained. It can adapt to pose estimation of objects with low texture, reflective surfaces, or partially occluded objects.

Description

technical field [0001] The invention relates to the technical field of machine vision, in particular to a pose estimation method and system of a target object, and a robot. Background technique [0002] In the field of robotics, autonomously grasping target objects is a key capability of intelligent robots. Among them, the grasping of scattered objects has always been the key to realizing the intelligence of robots. The ability of robots to grasp scattered objects can be applied to the classification of parts. Picking and other scenarios to improve work efficiency. But current robots can take weeks to reprogram to perform a complex new grasping task, making retooling modern manufacturing lines expensive and slow. Moreover, robots are mostly used in specific environments to perform grasping operations on specific known objects. For unknown objects placed in different poses in an uncertain environment, the robot can independently determine the grasping position and grasping p...

AI Technical Summary

Problems solved by technology

[0004] This application provides a pose estimation method and system of a target object, a robot and a computer-readable storage medium, aiming at solving the problem that most of the existing pose estimation methods rely on information such as the color and texture of the object surface for low texture, The problem of poor pose estimation of objects on reflective surfaces

Images