Scholarlink

Background <p>With advancing research in flexible robotics, human biomechanics and material failure mechanisms, higher requirements are imposed on measurement system adaptability. Traditional three-dimensional digital image correlation (3D-DIC) method employs fixed measurement position, limiting measurement range and position without adaptive adjustment capability.</p> Objective <p>This study aims to develop an adaptive 3D-DIC method capable of continuously tracking and measuring deformation of targets undergoing large-range both in-plane and out-of-plane motion, and pose variations.</p> Methods <p>The 3D-DIC system is affixed to the end effector of a robotic arm, which facilitates camera movement. This method achieves real-time tracking and observation of speckle targets through YOLOv11-based detection combined with position-based visual servoing. It also performs accurate deformation measurements by mapping the reconstructed 3D coordinates to a fixed world coordinate system to eliminate the effects of camera motion.</p> Results <p>The proposed system achieved sub-millimeter displacement accuracy, with maximum errors of 0.0558&#xa0;mm over 190&#xa0;mm in-plane motion and 0.0699&#xa0;mm over 165&#xa0;mm out-of-plane motion. Compared with fixed 3D-DIC, it produced more stable and accurate principal strain fields across large-range motions. Continuous observation was further validated under pose variations (20 poses, up to 105.2°). In finger–glass contact sliding experiments (197&#xa0;mm lateral and 136&#xa0;mm longitudinal), the method captured strain-field evolution from onset to steady sliding and localized stick–slip instants via velocity-spike detection.</p> Conclusions <p>The proposed robotic-assisted adaptive 3D-DIC method effectively addresses measurement challenges under complex motion conditions. This method effectively facilitates finger-glass contact measurements, providing more authentic experimental data for contact mechanics and haptics.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Robotic-Assisted Adaptive 3D Digital Image Correlation with Visual Servoing-Based Tracking

  • 期刊论文

  • 发表日期：

    2026年5月5日

查看全文
  • L. Huang,
  • X. Shao

摘要

Background

With advancing research in flexible robotics, human biomechanics and material failure mechanisms, higher requirements are imposed on measurement system adaptability. Traditional three-dimensional digital image correlation (3D-DIC) method employs fixed measurement position, limiting measurement range and position without adaptive adjustment capability.

Objective

This study aims to develop an adaptive 3D-DIC method capable of continuously tracking and measuring deformation of targets undergoing large-range both in-plane and out-of-plane motion, and pose variations.

Methods

The 3D-DIC system is affixed to the end effector of a robotic arm, which facilitates camera movement. This method achieves real-time tracking and observation of speckle targets through YOLOv11-based detection combined with position-based visual servoing. It also performs accurate deformation measurements by mapping the reconstructed 3D coordinates to a fixed world coordinate system to eliminate the effects of camera motion.

Results

The proposed system achieved sub-millimeter displacement accuracy, with maximum errors of 0.0558 mm over 190 mm in-plane motion and 0.0699 mm over 165 mm out-of-plane motion. Compared with fixed 3D-DIC, it produced more stable and accurate principal strain fields across large-range motions. Continuous observation was further validated under pose variations (20 poses, up to 105.2°). In finger–glass contact sliding experiments (197 mm lateral and 136 mm longitudinal), the method captured strain-field evolution from onset to steady sliding and localized stick–slip instants via velocity-spike detection.

Conclusions

The proposed robotic-assisted adaptive 3D-DIC method effectively addresses measurement challenges under complex motion conditions. This method effectively facilitates finger-glass contact measurements, providing more authentic experimental data for contact mechanics and haptics.