工件的位姿确定方法、装置、电子设备和程序产品

By solving the two-dimensional feature point set and three-dimensional model point set of the target workpiece, calculating the quality result and the number of interior points, and adaptively processing the pose, the problem that the PnP algorithm cannot evaluate the reliability is solved, and the reliability evaluation of the pose output is realized, ensuring the safety of industrial robot grasping and assembly.

CN122415718APending Publication Date: 2026-07-17LINGXIN QIAOSHOU (BEIJING) TECH CO LTD

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
LINGXIN QIAOSHOU (BEIJING) TECH CO LTD
Filing Date
2026-06-16
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

In existing technologies, the PnP algorithm cannot determine the reliability of the pose result, which leads to downstream systems receiving incorrect poses and executing incorrect actions, resulting in production accidents.

Method used

By solving the two-dimensional feature point set and three-dimensional model point set of the target workpiece, the initial pose and interior point set are obtained. Based on the number of interior points and reprojection error, the quality result is calculated, and the pose is adaptively processed. Output is only given when the pose is reliable; otherwise, the output is rejected.

Benefits of technology

This ensures the safety of industrial robots in grasping and assembling, and avoids production accidents caused by incorrect positioning.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122415718A_ABST
    Figure CN122415718A_ABST
Patent Text Reader

Abstract

本申请公开了一种工件的位姿确定方法、装置、电子设备和程序产品,属于工业检测技术领域。所述工件的位姿确定方法包括:对获取的目标工件对应的二维特征点集和三维模型点集进行求解,得到所述目标工件的初始位姿和内点集合;基于所述内点集合中的内点数量对所述初始位姿进行处理,得到所述目标工件的目标位姿;基于所述目标位姿的质量结果,输出所述目标位姿或拒绝输出所述目标位姿。本申请的工件的位姿确定方法使位姿求解过程具备了评估自身输出可信度的能力,只有在位姿可信时才输出,不可信时主动拒绝输出,从而避免了下游系统因接收错误位姿而导致的生产事故,保障了工业机器人抓取和装配的安全性。
Need to check novelty before this filing date? Find Prior Art