丁超, 唐力伟, 曹立军, 邵新杰, 邓士杰. 复杂深孔内轮廓结构光图像畸变矫正算法[J]. 红外与激光工程, 2017, 46(12): 1217008-1217008(7). DOI: 10.3788/IRLA201746.1217008
引用本文: 丁超, 唐力伟, 曹立军, 邵新杰, 邓士杰. 复杂深孔内轮廓结构光图像畸变矫正算法[J]. 红外与激光工程, 2017, 46(12): 1217008-1217008(7). DOI: 10.3788/IRLA201746.1217008
Ding Chao, Tang Liwei, Cao Lijun, Shao Xinjie, Deng Shijie. Image distortion correction algorithm for complicated deep-hole profile using structured-light[J]. Infrared and Laser Engineering, 2017, 46(12): 1217008-1217008(7). DOI: 10.3788/IRLA201746.1217008
Citation: Ding Chao, Tang Liwei, Cao Lijun, Shao Xinjie, Deng Shijie. Image distortion correction algorithm for complicated deep-hole profile using structured-light[J]. Infrared and Laser Engineering, 2017, 46(12): 1217008-1217008(7). DOI: 10.3788/IRLA201746.1217008

复杂深孔内轮廓结构光图像畸变矫正算法

Image distortion correction algorithm for complicated deep-hole profile using structured-light

  • 摘要: 为实现针对复杂深孔内轮廓表面几何参数的高精度测量,建立了基于结构光的内轮廓检测系统,但由于深孔内轮廓的曲面特性,相机采集到的图像相比于内轮廓展开后的平面图像存在较大的几何畸变,直接影响到内壁几何参数的量化精度。首先针对深孔内轮廓和内轮廓展开结构进行无差别建模,分析两者之间的空间坐标转换关系;然后综合考虑畸变矫正的精度和速度,借鉴三次样条插值函数、离散映射等理念提出了基于离散映射的深孔内轮廓矫正算法,实现针对深孔内轮廓的畸变在线矫正,检测精度达到亚像素水平,低于0.1 mm。

     

    Abstract: To achieve the high precision measurement of the complicated deep-hole profile geometric parameters, the detection system for the profile based on the structured light was established. However the images collected by the camera had larger geometric distortion compared with the plane images from the profile because of the curvy surface features of the deep-hole profile. The geometric distortion affected the calculation accuracy of the geometric parameters directly. Firstly, the models were established indiscriminately for the deep-hole profile and its plane structure to analyze the space coordinate transformation between each other. Then, the ideas of the cubic spline interpolation and discrete mapping were referred to propose the deep-hole profile correction algorithm based on the discrete mapping considering the precision and speed of the distortion correction comprehensively. The purpose of correcting the deep-hole profile distortion online was achieved through the algorithm. The test precision reached the sub pixel level and the error was less than 0.1 mm.

     

/

返回文章
返回