期刊文献+

尺度各异刚体碎块断裂面的匹配方法 被引量:1

A Matching Method Based on Fracture Surface for Rigid Blocks with Different Scale
下载PDF
导出
摘要 针对厚度不可忽略的刚体碎块,提出一种尺度刚体碎块匹配方法,以解决碎块匹配中的刚体变换和尺度变换的问题.采用改进的区域生长算法对碎块的外表面进行分割,并根据曲面的粗糙程度提取碎块的断裂面;通过添加尺度矩阵、旋转角约束和动态迭代系数的方式来改进迭代最近点(iterative closest point,ICP)算法,并采用该改进的ICP算法实现碎块的断裂面匹配.实验结果表明,跟ICP算法相比,改进的ICP算法不仅能够克服ICP算法不能解决的尺度变换问题,而且与尺度ICP(SICP)算法相比,可以更加精确、快速地实现碎块的断裂面匹配. Aiming at rigid blocks with a certain thickness,a scaling rigid block matching method was proposed to solve the problems of rigid transformation and scale transformation in block matching process.At first,the surface of each block was segmented by an improved region growing algorithm,and the fracture surfaces of the block were extracted according to its roughness.Then,an improved iterative closest point(ICP)algorithm was proposed by integrating scale matrix,rotation angle constraint and active iterative coefficient to ICP algorithm.Finally,the improved ICP algorithm was used to complete fracture surfaces matching of rigid blocks.The experimental results show that,compared with ICP algorithm,the improved ICP algorithm can overcome the problem of scaling transformation;And compared with scaling ICP(SICP)algorithm,the improved ICP algorithm can get much higher accuracy and convergence rate in fracture surface matching process.
作者 赵夫群 周明全 耿国华 ZHAO Fu-qun;ZHOU Ming-quan;GENG Guo-hua(College of Information Science and Technology,Northwest University,Xi'an,Shannxi 710127,China;College of Education Science,Xianyang Normal University,Xianyang,Shannxi 712000,China;College of Information Science and Technology,Beijing Normal University,Beijing 100875,China)
出处 《北京理工大学学报》 EI CAS CSCD 北大核心 2019年第1期95-100,共6页 Transactions of Beijing Institute of Technology
基金 国家自然科学基金资助项目(61731015 61672013)
关键词 碎块匹配 迭代最近点 尺度矩阵 旋转角约束 动态迭代系数 blocks matching iterative closest point scale matrix rotation angle constraint active iterative coefficient
  • 相关文献

参考文献8

二级参考文献51

共引文献96

同被引文献14

引证文献1

二级引证文献3

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部