摘要
光栅式双目立体视觉传感器用于物体三维形貌测量,其难点之一是光栅光条的匹配与识别.利用可见光条数和光条序列组合间的关系,提出了一种基于空间三维搜索的快速光条匹配算法.将光栅式双目立体视觉传感器看作两个光栅结构光传感器,分别进行标定后,可计算各光条中心点的三维坐标.搜索不同光条中心点集之间的距离最小且模型编号相同的光条即为匹配光条.分析了可见光条数目和光条序列组合间的关系后,列出各种可能的组合情况,设计了搜索算法,排除了冗余计算,大大减少了计算量,实验证明了方法的有效性.
Raster binocular stereo vision sensor is widely used in 3D measurement,and one of its difficulties is light-stripe matching and recognizing.Based on the relationship between the viewed stripes number and the stripe sequence combination,a fast matching method for light-stripe based on 3D searching was proposed.The raster binocular stereo vision system was regarded as two raster structured light sensors,and then was calibrated under one common coordinated system.So,the 3D coordinates of the light stripes center points were obtained by the two raster structured light sensors,respectively,as well as the distances between the light-stripe center points.As a result,the light-stripes whose model numbers are equal and have the minimal distance are matching stripes.All the possible combinations were listed after the relationship between the viewed stripes number and the stripe sequence combination,and the matching method was designed with the character of much less extra calculations which brings about less computational burden.The experiment results show the effectiveness of the method.
出处
《北京航空航天大学学报》
EI
CAS
CSCD
北大核心
2012年第5期591-594,600,共5页
Journal of Beijing University of Aeronautics and Astronautics
基金
国家自然科学基金资助项目(50875014)
北京市自然科学基金资助项目(3092014)
关键词
光条匹配
立体视觉
光栅
light-stripes matching
stereo vision
raster
