【作者】 严雪梅； 戴宁； 崔海华； 程筱胜；

【Author】 Yan xue-mei Dai Ning Cui Hai-hua Cheng Xiao-sheng (1 college of Mechanical and Electrical Engineering, Nanjing University of Aeronautics & Astronautics,Nanjing, Jiangsu,210016,china)

【机构】 南京航空航天大学机电学院；

【摘要】 提出一种基于单幅彩色图像的编码匹配方法。首先在色彩编码结点的有向图中进行哈密顿回路搜索,利用当前条纹颜色及其相邻三个条纹颜色确定编码条纹,将被测量物体的图像空间在一维方向上分成若干唯一的编码区域,结合双目立体视觉的极线约束实现区域级粗匹配;然后利用数字图像相关性算法,在粗匹配的局部区域内实现像素级致密精匹配。实验验证,提出的方法能精确稳定的实现用于三维测量的单幅彩色图像的编码匹配。更多还原

【Abstract】 This paper proposes a matching method which is based on a single color image.First,the color code was produced by searching a Hamiltonian Circuit in a directed graph of color coding nodes.The current color stripe and the adjacent three stripes decided the uniqueness position of the stripe.The stripes divided the object image into several uniqueness parts in the horizontal direction.Combined with the epipole constrain of binocular stereo vision,the regional coarse matching was realized.Then,using the correlation matching algorithm of digital images,dense matching was realized.Experiment results show that the proposed method is accurate and robust for the matching of 3D measurement.更多还原