Abstract: A data registration method based on planar baseline target is proposed for the 3D data registration problem in vision measurement. An optimization object function about all parameters of rotation matrix and transmission vector between two vision sensor coordinate frames at adjacent viewpoints is established using feature points provided by the planar target. The rotation matrix and the transmission vector with an unknown scale are estimated by Levenberg-Marquardt optimiz- ation algorithm. The unknown scale is found relying on base- line length on the planar target, and the data registration matrix is obtained. The registration experiment is conducted using 100 feature points on the planar target, and the data registration RMS error of x, y and z is 0.038 mm, 0.022 mm and 0.135 mm respectively. The experimental results show the presented method is flexible and widely applicable.

Key words: Planar target, 3D data registration, Baseline length, Vision measement