Flexible vision shape measurement by multi-way combination intersection

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.