A Multi-view Measurement Algorithm Based on Bundle Adjustment for the Position of Optical Fibers

摘要

The multi-view vision measurement method has been widely used due to its high efficiency and high-precision. But, the general vision measurement method is not applicable to many occasions, such as the position measurement of the optical fibers in the astronomical telescope. In this paper, a novel vision measurement method is proposed based on the characteristics of optical fiber. In order to get higher measurement accuracy in large field working space, precise feature point extraction and improved bundle adjustment are researched. Firstly, the vision measurement method and the measuring process are introduced. Then, the camera model and calibration are presented. Finally, the improved measurement algorithm is investigated in detail based on iterative bundle adjustment and additional constraints which includes measuring points and only one high-precision control point. The experimental results indicate that the new approach is more accurate and more reliable than the direct vision measurement. The maximum position error of the novel measurement algorithm of optical fibers with real data is 24 μm and all the position errors conform to the accuracy requirement.