摘要
针对图像配准经典迭代最近点(ICP,iterative closest points)算法存在的收敛效率低、容易陷入局部收敛的问题,给出了一种基于掌纹图像识别的改进ICP算法。该算法结合经典ICP算法和掌纹识别算法的优势,同时准确地嵌入容器模板类来存储图像的像素坐标信息,并引入二元搜索树(k-d tree)来寻找最近点集合,最终得到一个基于香港理工大学公开掌纹图像库的图像配准程序系统。实验表明:通过该系统所得到的结果具有较好的鲁棒性和适用性。
Image registration is playing an important role in the computer image processing.Many investigators had pay close attention to the ICP algorithm which is widely used to solve the problem of registration.As the classical ICP algorithm has poor efficiency and is easy to fall into local convergence,this paper will put forward an improved ICP algorithm,The algorithm combines the advantages of classical ICP algorithm and palmprint recognition algorithm,meanwhile it had introduced k-d tree to find the nearest point set and used container class templates to solve storage issues,at last,it had given out a program system based on PolyU Palmprint Database and can do well in image registration with good robustness and applicability.
出处
《北京信息科技大学学报（自然科学版）》
2013年第3期49-52,56,共5页
Journal of Beijing Information Science and Technology University
基金
北京市自然科学基金资助项目(1072010)
关键词
迭代最近点
图像配准
旋转矩阵
平移向量
iterative closest points
image registration
rotation matrix
translation vector
