A flexible 3D scanning system using convergent stereo vision integrated with an industrial manipulator

被引：0

作者：

Thanh Nhat Luong ^{[1
]}

Vo Hoang Hung Do ^{[2
]}

Quoc Chi Nguyen ^{[2
,3
]}

机构：

[1] Think Alpha Co, Dept Res & Dev, Ho Chi Minh City, Vietnam

[2] Ho Chi Minh City Univ Tech, Dept Mechatron, Ho Chi Minh City, Vietnam

[3] Ho Chi Minh City Univ Tech, DCSELab, Ho Chi Minh City, Vietnam

来源：

2021 INTERNATIONAL CONFERENCE ON HIGH PERFORMANCE BIG DATA AND INTELLIGENT SYSTEMS (HPBD&IS) | 2021年

关键词：

3D scanning; toe-in stereo vision; point cloud registration; point cloud processing; EYE-TO-HAND; PROJECTION;

D O I：

10.1109/HPBDIS53214.2021.9658437

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This paper proposes a flexible three-dimensional (3D) scanning system based on an active stereo scanner mounted on a six degrees of freedom (6-DOF) industrial robot arm. In addition, we reduce the blind area of sensors by adding one more DOF to adjust the convergent angle of the two cameras. This configuration is suitable for scanning items with different sizes, complex shapes, and sophisticated structures due to the flexibility in approaching the object. Then, the research develops a robot-based iterative closest point method for point cloud registration. This 3D reconstruction strategy enhances the accuracy as well as the computational time of the point cloud processing. After that, we conduct several real-world experiments to examine the effectiveness of the proposed system and algorithm.

引用

页码：167 / 171

页数：5