基于多向联合交汇的柔性视觉形貌测量新方法

Flexible vision shape measurement by multi-way combination intersection

  • 摘要: 针对大尺寸物体形貌测量应用,设计一种视觉形貌测量新方法。该方法是将两套没有公共视场的单目视觉传感器组合起来联合交汇,分别用于位姿测量和单目多位置交汇测量。位姿测量通过求解PnP问题为单目多位置交汇测量提供位姿信息。单目多位置交汇测量利用位姿测量提供的辅助信息通过光束平差解算出待测物的形貌。分析了系统原理,给出位姿测量PnP算法实现过程,介绍了非公共视场相机的标定方法原理,分析了单目多位置交汇测量的原理并给出实现方法。最后进行了测量方法的验证实验,实验结果表明,在3 m×3 m的工作范围内系统测量精度为5 mm,证明该方法可以用于大尺寸物体形貌测量,且结构简单轻便,具有很强的灵活性。

     

    Abstract: This paper describes new vision measurement method for shape measurement of large-size object. Combine and intersect two monocular vision measurement systems without public filed of view for measurements of pose estimation and monocular vision intersection respectively. Pose estimation system solves the PnP problem to provide position information for monocular measurement. Monocular measurement system combining with auxiliary information from pose estimation calculates shape information of object. This paper analyses the system principle, gives PnP algorithm's implementation process of pose estimation system, and introduces the principle and implementation process of calibration method for camera without public field of view. Finally, static system verification test of measurement system of indoor large-size object has been carried out, and the feasibilities of this system have been validated. Experimental results show that measuring accuracy of system is 5 mm in the scope of the 3 m×3 m. It is proved that this system can be used in shape measurement of large-size object, and the structure of system is simple and light with great flexibility.

     

/

返回文章
返回