基于轴线投影精确模型的弯管立体视觉测量方法
|
摘要
利用立体视觉技术测量弯管轴线需精确计算弯管轴线在图像上的投影位置,现有方法将弯管图像轮廓中心线近似为空间轴线在像平面上的投影,这种近似计算存在误差。为获得弯管轴线的精确投影位置,提高三维测量精度,对弯管成像特征进行了深入研究。首先对弯管任意横截面建立透视投影模型,探索截面图像边缘点与空间轴线投影点间的几何关系,进而提出轴线上任意一点在像平面上投影位置的精确计算方法。采用双目视觉系统对某一航空管件进行重构测量。试验结果表明,相较于现有方法,采用本文方法计算轴线投影位置后,管件测量的重复性精度与前者相当,而测量精度平均提高14.5%。说明本文方法对轴线投影位置计算正确,可有效提高弯管轴线视觉测量精度。
The projection position of tube axis should be calculated accurately when the tube axis is measured using stereo vision. Previous methods for measuring tube axis approximate the centerlines of image contours as the axis projection. This assumption exists errors. In order to obtain the exact projection position of tube axis and to improve the measurement precision of tube axis, the proposed method studied the imaging characteristics of tubes. Firstly, a perspective projection model for any cross section of tube was established to explore the geometric relationship between the contour points of the cross section and their corresponding projection points on space axis. Furthermore, a way to locate the precise projected position on image planes of points laying on axis was presented. A binocular vision system was constructed to measure an aerospace tube. Compared with classical approaches, the repeatability accuracy of the proposed method is same as the former, while the measurement accuracy of the presented method increased by 14.5% on average due to precisely calculating the projection position of axis. It is shown that the proposed method is correct in calculating the projection position of tube axis and can effectively improve the measurement precision of tube axis.
The projection position of tube axis should be calculated accurately when the tube axis is measured using stereo vision. Previous methods for measuring tube axis approximate the centerlines of image contours as the axis projection. This assumption exists errors. In order to obtain the exact projection position of tube axis and to improve the measurement precision of tube axis, the proposed method studied the imaging characteristics of tubes. Firstly, a perspective projection model for any cross section of tube was established to explore the geometric relationship between the contour points of the cross section and their corresponding projection points on space axis. Furthermore, a way to locate the precise projected position on image planes of points laying on axis was presented. A binocular vision system was constructed to measure an aerospace tube. Compared with classical approaches, the repeatability accuracy of the proposed method is same as the former, while the measurement accuracy of the presented method increased by 14.5% on average due to precisely calculating the projection position of axis. It is shown that the proposed method is correct in calculating the projection position of tube axis and can effectively improve the measurement precision of tube axis.
引文
[1]孔筱芳,陈钱,顾国华,等,GPS双目摄像机标定及空间坐标重建[J]·光学精密工程,2017,25(2):485-492-KONG Xiaofang, CHEN Qian, GU Guohua, et al·GPS-based binocular camera calibration and space coordinate reconstruction[J]·Optics and Precision Engineering, 2017,25(2):485-492.
[2]倪章松,顾艺,柳庆林,等·大视场双目立体视觉柔性标定[J]·光学精密工程,2017, 25(7):1882-1889·NI Zhangsong, GU Yi, LIU Qinglin, et al·Flexible calibration method for binocular stereo vision in large field of view[J]·Optics and Precision Engineering,2017, 25(7):1882-1889·
[3]杨景豪,刘巍,刘阳,等·双目立体视觉测量系统的标定[J]·光学精密工程,2016, 24(2):300-308.YANG Jinghao, LIU Wei, LIU Yang, et al·Calibration of binocular vision measurement system[J]·Optics and Precision Engineering,2016, 24(2):300-308.
[4] BOSEMANN W·Industrial photogrammetry. accepted metrology tool or exotic niche[J]·ISPRS-International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2016, 5:15-24·
[5] JIN P,LIU J H,LIU S,et al·Automatic multi-stereo-vision reconstruction method of complicated tubes for industrial assembly[J]·Assembly Automation, 2016,36(4):362-375.
[6] LIU S, JIN P, LIU J, et al·Accurate measurement method for tube's endpoints based on machine vision[J]·Chinese Journal of Mechanical Engineering, 2017, 30(1):1-12.
[7] LIU S, LIU J, JIN P, et al·Tube measurement based on stereo-vision:a review[J]. International Journal of Advanced Manufacturing Technology,2017, 92(4):1-16.
[8] ZHANG T,LIU J,LIU S,et al·A 3D reconstruction method for pipeline inspection based on multi-vision[J]·Measurement, 2017,98:35-48.
[9] DOIGNON C,DE MATHELIN M·A degenerate conic-based method for a direct fitting and 3-D pose of cylinders with a single perspective view[C]//IEEE International Conference on Robotics and Automation.New York:IEEE, 2007·
[10] LIU C, HU W·Relative pose estimation for cylinder-shaped spacecrafts using single image[J]·IEEE Transactions on Aerospace&Electronic Systems,2014,50(4):3036-3056·
[11] BOSEMANN W·The optical tube measurement system OLM photogrammetric methods used for industrial automation and process control[J]·International Archives of Photogrammetry and Remote Sensing, 1996, 31:55-58·
[12] LEE I K, KIM K J·Shrinking:Another method for surface reconstruction[C]//Geometric Modeling and Processing·Washington:IEEE Computer Society, 2004·
[13] CAGLIOTI V, GIUSTI A·Reconstruction of canal surfaces from single images under exact perspective[C]//Proceedings of Computer Vision-ECCV 2006, European Conference on Computer Vision·Graz:DBLP,2006.
[14] TRUJILLO-PINO A, KRISSIAN K,ALEMAN-FLORES M, et al·Accurate subpixel edge location based on partial area effect[J]·Image&Vision Computing,2013, 31(1):72-90·
[2]倪章松,顾艺,柳庆林,等·大视场双目立体视觉柔性标定[J]·光学精密工程,2017, 25(7):1882-1889·NI Zhangsong, GU Yi, LIU Qinglin, et al·Flexible calibration method for binocular stereo vision in large field of view[J]·Optics and Precision Engineering,2017, 25(7):1882-1889·
[3]杨景豪,刘巍,刘阳,等·双目立体视觉测量系统的标定[J]·光学精密工程,2016, 24(2):300-308.YANG Jinghao, LIU Wei, LIU Yang, et al·Calibration of binocular vision measurement system[J]·Optics and Precision Engineering,2016, 24(2):300-308.
[4] BOSEMANN W·Industrial photogrammetry. accepted metrology tool or exotic niche[J]·ISPRS-International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2016, 5:15-24·
[5] JIN P,LIU J H,LIU S,et al·Automatic multi-stereo-vision reconstruction method of complicated tubes for industrial assembly[J]·Assembly Automation, 2016,36(4):362-375.
[6] LIU S, JIN P, LIU J, et al·Accurate measurement method for tube's endpoints based on machine vision[J]·Chinese Journal of Mechanical Engineering, 2017, 30(1):1-12.
[7] LIU S, LIU J, JIN P, et al·Tube measurement based on stereo-vision:a review[J]. International Journal of Advanced Manufacturing Technology,2017, 92(4):1-16.
[8] ZHANG T,LIU J,LIU S,et al·A 3D reconstruction method for pipeline inspection based on multi-vision[J]·Measurement, 2017,98:35-48.
[9] DOIGNON C,DE MATHELIN M·A degenerate conic-based method for a direct fitting and 3-D pose of cylinders with a single perspective view[C]//IEEE International Conference on Robotics and Automation.New York:IEEE, 2007·
[10] LIU C, HU W·Relative pose estimation for cylinder-shaped spacecrafts using single image[J]·IEEE Transactions on Aerospace&Electronic Systems,2014,50(4):3036-3056·
[11] BOSEMANN W·The optical tube measurement system OLM photogrammetric methods used for industrial automation and process control[J]·International Archives of Photogrammetry and Remote Sensing, 1996, 31:55-58·
[12] LEE I K, KIM K J·Shrinking:Another method for surface reconstruction[C]//Geometric Modeling and Processing·Washington:IEEE Computer Society, 2004·
[13] CAGLIOTI V, GIUSTI A·Reconstruction of canal surfaces from single images under exact perspective[C]//Proceedings of Computer Vision-ECCV 2006, European Conference on Computer Vision·Graz:DBLP,2006.
[14] TRUJILLO-PINO A, KRISSIAN K,ALEMAN-FLORES M, et al·Accurate subpixel edge location based on partial area effect[J]·Image&Vision Computing,2013, 31(1):72-90·