工程设计学报2025，Vol.32Issue(5)：613-622,10.DOI:10.3785/j.issn.1006-754X.2025.04.156
基于双目视觉的软体机械臂几何参数测量方法
Geometric parameter measurement method for soft robotic arms based on binocular vision
摘要
Abstract
Due to the inherent compliance and low stiffness of soft robotic arms,they exhibit complex morphological changes during spatial motion,making existing measurement methods inadequate for pose measurement of such robotic arms.To solve the above problems,a novel method combining binocular vision technology and B-spline curve fitting is proposed to accurately measure the geometric parameters of soft robotic arms.This method extracted the three-dimensional information of the central skeleton of the soft robotic arm through a binocular vision system and employed the cubic B-spline curve for three-dimensional shape reconstruction,thereby obtaining the key geometric parameters of the soft robotic arm.To validate the effectiveness of the proposed method,geometric parameter measurements were conducted on soft robotic arms with arc-shaped,S-shaped and L-shaped bends.The measurement results showed that the average bending curvature error of the robotic arm moving on a plane was 0.198%,and the average bending angle error was 0.159%.The average bending angle error of the robotic arm moving in three-dimensional space was 1.340%.In addition,this measurement method could be extended to dynamic measurements and accurately measure the geometric parameters of soft robotic arms even in the presence of visual occlusions.The measurement method based on binocular vision and B-spline curve fitting can provide a new idea for the parameter measurement of soft robotic arms.关键词
软体机械臂/双目视觉/B样条曲线/几何参数测量Key words
soft robotic arm/binocular vision/B-spline curve/geometric parameter measurement分类
计算机与自动化引用本文复制引用
舒申,王家梁,胡俊峰,张宇,楚凯,周浩,蔡铭炜..基于双目视觉的软体机械臂几何参数测量方法[J].工程设计学报,2025,32(5):613-622,10.基金项目
国家自然科学基金资助项目(52165011) （52165011）
江西省自然科学基金资助项目(20212BAB204028) （20212BAB204028）
