单目结构光在异形工件三维重构及测量中的应用
3D reconstruction measurement system of irregular workpieces based on monocular structured light
王志强 1任文佳 1乔什 1李跃华 2张桂英3
作者信息
- 1. 天津职业技术师范大学信息技术工程学院,天津 300222
- 2. 天津市微低重力环境模拟技术重点实验室,天津 300301
- 3. 天津职业技术师范大学电子技术工程学院,天津 300222
- 折叠
摘要
针对工业生产中异形工件尺寸在线测量的难题,将单目相机与线激光相结合,以突破传统测量技术在复杂几何形状、成本、避免表面损伤和测量精度方面的局限性.针对线结构光条纹中心提取的挑战,对传统的Steger算法改进提出双重调节优化算法,优化Steger算法中的二值化阈值参数和光条宽度参数,以提高算法在检测条纹中心时的准确性和鲁棒性.实验结果表明:采用基于线结构光和单目相机的测量方法以及改进的结构光条纹中心提取技术,对异形工件进行了三维重构,测量误差在 0.10 mm以内,处于可接受的范围.相较于传统方法,基于线结构光和单目相机的新测量方法提取的条纹中心位置更准确,能实现对异形工件的高精度测量.
Abstract
This study addresses the challenge of accurately measuring the dimensions of irregular workpieces in industrial production by integrating a monocular camera with line laser technology.This innovative approach overcomes the limitations of traditional measurement techniques,which often struggle with complex geometries,cost considerations,the need to avoid surface damage,and achieving high precision.To improve the extraction of the stripe center of line structured light,a dual-parameter optimization algorithm is proposed to enhance the traditional Steger algorithm by fine-tuning binary threshold pa-rameters and line width parameters,thereby significantly improving the accuracy and robustness of stripe center detection.Experimental findings demonstrate the effectiveness of the line structured light measurement method combined with a monocular camera,achieving successful 3D reconstruction of irregular workpieces with a measurement error of less than 0.10 mm,well within acceptable limits.Compared to conventional methods,this new measurement technique offers superior accuracy in stripe center extraction of line structured light,facilitating high-precision measurements of irregular workpieces while preventing surface damage and remaining cost-effective.
关键词
异形工件/线结构光/条纹中心提取/三维重构/三维测量Key words
irregular workpieces/line structured light/stripe center extraction/3D reconstruction/3D measurement引用本文复制引用
基金项目
天津市教委科研计划项目(2022KJ118)
天津职业技术师范大学科研启动项目(KYQD1625)
出版年
2024
NETL
NSTL
万方数据