Scholarlink

<p>In modern industrial and scientific research fields, the demand for high-precision 3 Dimensions (3D) measurement of non-Lambertian complex surfaces is increasing, but traditional measurement techniques perform poorly in terms of accuracy, efficiency, and applicability. To this end, an innovative fusion of structured light stripes, light decoding technology, and improved stereo matching algorithm was studied, and a 3D measurement system based on dual eye stereo vision was proposed. The system obtains phase information through encoding structured light stripe pattern projection and phase shift method, and uses multi-frequency heterodyne method to expand phase. The results showed that in the JanusGraph dataset, the accuracy of the improved algorithm was close to 100%, while in other datasets, its accuracy exceeded 90%. And the improved phase shift method had lower errors in processing complex surfaces, with errors increasing from about 0.005–0.02, far lower than other methods. The experimental results demonstrate the high accuracy and stability of the proposed method for measuring complex surfaces. The research results contribute to improving the accuracy of traditional measurement techniques and promoting the development of modern industry and scientific research.</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

3D Measurement Method for Non-Lambertian Complex Surfaces Based on Stereo Vision

  • 期刊论文

  • 发表日期：

    2025年7月3日

查看全文
  • Ruihua Zhang,
  • ShuBo Bi

摘要

In modern industrial and scientific research fields, the demand for high-precision 3 Dimensions (3D) measurement of non-Lambertian complex surfaces is increasing, but traditional measurement techniques perform poorly in terms of accuracy, efficiency, and applicability. To this end, an innovative fusion of structured light stripes, light decoding technology, and improved stereo matching algorithm was studied, and a 3D measurement system based on dual eye stereo vision was proposed. The system obtains phase information through encoding structured light stripe pattern projection and phase shift method, and uses multi-frequency heterodyne method to expand phase. The results showed that in the JanusGraph dataset, the accuracy of the improved algorithm was close to 100%, while in other datasets, its accuracy exceeded 90%. And the improved phase shift method had lower errors in processing complex surfaces, with errors increasing from about 0.005–0.02, far lower than other methods. The experimental results demonstrate the high accuracy and stability of the proposed method for measuring complex surfaces. The research results contribute to improving the accuracy of traditional measurement techniques and promoting the development of modern industry and scientific research.