基于组合编码的条纹结构光主动立体视觉匹配

Automatic Acquiring of Measurement Points for CMM Under CATIA V5 Based on The Digital Model of Product[HS)]

摘要:

立体视觉中如何提高图像匹配的密度以及精度,已成为研究中的热点和难点；提出Gray码条纹和正弦位移条纹投射相结合组合编码方案,结合中心线解码完成图像匹配,可消除Gray码解码误差,以及基于绝对相位匹配中的相位展开误差；位移步长不受条纹周期的限制,显著提高了采样点的匹配密度；采用双向正交投射实现视场全局唯一性编码,作为匹配特征完成同名采样点的一一对应；应用该编解码方案建立主动视觉测量系统,完成了标准平板和圆柱的形状检测与误差评定,以及复杂形状物体的三维重构；结果表明了算法的有效性。

Abstract:

The question that how to increase the density and the accuracy of image matching in stereoscopic vision has become a hotspot and difficulty of research. A kind of combination coding scheme, which is based on the projection of fringes of Gray code and sine displacement, is proposed in this paper. In this way, the image matching is carried out by means of center line decoding method, which results in removing of decoding error derived in Gray code method, and avoiding of the phase expanding error generated in absolute phase based image matching. Due to the fact that the displacement step is not restricted by the period of the fringe, the matching density of image sample points can be increased significantly. The projecting of combination fringes of two orthogonal directions achieves the goal of global uniquely coding within the whole view field, which is taken as the matching factor to search the corresponding sample points. Measuring experiments of a number of target objects with various sizes and shapes indicate the validity of the method proposed.