基于遗传算法的三维曲面配准  

Three Dimensional Surface Registration on the Basis of Genetic Algorithm

作　　者：刘纯国[1] 刘畅[1] 安百玲 

机构地区：[1]吉林大学辊锻研究所,吉林长春130022 [2]吉林省石油化工设计研究院,吉林长春130022

出　　处：《锻压装备与制造技术》2009年第4期110-113,共4页China Metalforming Equipment & Manufacturing Technology

基　　金：国家自然科学基金资助项目(50275063);教育部留学回国人员科研启动基金项目

摘　　要：本文针对自由曲面模型数据和测量数据之间的配准问题,研究了先粗后精的两步配准方法。利用遗传算法的全局搜索性能,控制优化参数达到两曲面间的粗配准;以粗配准的结果作为初值,利用ICP匹配算法修正误差,进一步优化以达到全局最优配准。两步法解决了ICP算法易收敛于局部最小值的问题,将测量曲面与CAD模型曲面在空间进行最优化匹配。实例验证表明,算法配准精度高、运算速度比较快,可以应用于曲面件多点成形后的曲面误差分析及修正,也可应用于其他领域自由光滑曲面的测量结果的比较分析。In order to calculate the shape error between the formed work-piece and its CAD model in reverse engineering, two-step surface registration method combining ICP algorithm and Genetic Algorithm optimization has been proposed. With Genetic Algorithm method, the initial surface error has been calculated by aligning the data from different coordinate system in first rough registration step.ICP algorithm method has been employed to calculate the surface error further in second step.The global optimal registration result can be obtained while the local convergence has been avoided in the algorithm. Due to its rapid convergence characteristic, it is not only used in analysis and amending of the curved surface error after multi-point forming, but＇ also in the comparison between measure results of different smooth curved surface in other industry.

关 键 词：计算机应用 配准 三维曲面 遗传算法 

分 类 号：TG391.68[金属学及工艺—金属压力加工]