线结构光三维扫描测量中的多视数据对齐

Alignment of multi-view data of linear structure light scanning measurement

对于大尺寸复杂形状全貌测量,基于光学测量技术测量范围的限制,需要将多视角测量数据统一到一个全局坐标系下,以形成被测形状的完整表达。该文研究了线结构光测量环境下的多视数据对齐技术,提出以投射激光光条为依据的标志点识别与计算方法,应用距离约束法建立各视角下标志点集的对应关系。利用各视角下的标志点集建立局部测量坐标系,并利用局部测量坐标系实现多视测量数据的预对齐;而且以测量数据的预对齐作为初始对齐结果,使用优化算法进行多视数据精确对齐以提高对齐精度。实验结果表明,多视数据对齐方法快速、稳定,可有效扩展线结构光测量技术的测量范围。

Restricted by measurement range, measurement methods based on optical technology need to transform multiview data into a global coordinate system to form complete expression of body being measured with large size and complicated shape. This paper focused on alignment of multiview data and put forward recognition and calculation method of marked points from laser strips projected by laser sensor, as well as matching of marked point sets of different views with distance constraint method. Lo cal measurement coordinate system was established by marked points of each view, which was applied to achieve pre^alignment of multi-view data. ICP method combined with the pre-alignment method was used to accurately locate data of different views. Ex perimental results showed that the method was rapid and stable, and could effectively extend measurement scope of linear struc- ture light method.