|
|||||
|
|
| 面向地图点集的多尺度层级ICP算法 | |
| 引用本文： | 李策,罗欣颖,杜少毅.面向地图点集的多尺度层级ICP算法[J].计算机应用与软件,2012,29(4):83-85,93. |
| 作者姓名： | 李策 罗欣颖 杜少毅 |
| 作者单位： | 1. 兰州理工大学电气工程与信息工程学院 甘肃兰州730050;西安交通大学人工智能与机器人研究所 陕西西安710049;甘肃省工业过程先进控制重点实验室 甘肃兰州730050 2. 兰州理工大学电气工程与信息工程学院 甘肃兰州730050;甘肃省工业过程先进控制重点实验室 甘肃兰州730050 3. 甘肃省工业过程先进控制重点实验室 甘肃兰州730050 |
| 基金项目： | 国家自然科学青年基金项目(61005014);甘肃省自然科学基金项目(1112RJZA033) |
| 摘 要： | 地图点集具有点数多、结构复杂等特点,通常对其配准耗时严重,难以满足自主驾驶等情况下的实时性要求.利用多尺度层级化思想,提出一种多尺度层级ICP算法MSICP( Multi-scale Iterative Closest Points),提高了配准速度和精度.所提算法先对待配准图像点集进行稀疏化,随后将稀疏点集配准后的转换矩阵作为原稠密点集配准的转换矩阵初始值,最终实现对原始图像点集的ICP快速精确配准.实验结果表明,所提算法的配准速度及精度优于其他ICP算法,具有一定的实用价值. |
| 关 键 词： | 地图点集 ICP算法 多尺度 |
A MAP POINT SETS-ORIENTED MULTI-SCALE ICP ALGORITHM |
|
| Li Ce , Luo Xinying , Du Shaoyi.A MAP POINT SETS-ORIENTED MULTI-SCALE ICP ALGORITHM[J].Computer Applications and Software,2012,29(4):83-85,93. | |
| Authors: | Li Ce Luo Xinying Du Shaoyi |
| Affiliation: | 1(College of Electrical and Information Engineering,Lanzhou University of Technology,Lanzhou 730050,Gansu,China) 2(Institute of Artificial Intelligence and Robotics,Xi’an Jiaotong University,Xi’an 710049,Shaanxi,China) 3(Key Laboratory of Gansu Advanced Control for Industrial Processes,Lanzhou 730050,Gansu,China) |
| Abstract: | Map point sets have some features such as multiple points and complex structure,etc.,and their registration is usually time-consuming and can hardly meet the real-time requirements in circumstances such as autonomous driving and so on.In this paper we propose a multi-scale iterative closest points algorithm(MSICP) by using the idea of multi-scale hierarchy,which improves the speed and accuracy of registration.In the proposed algorithm,first the graphics point sets to be registered are made sparseness;then the translation matrix of sparse point set that has been matched is employed as initial values of translation matrix of original dense point set;and at last the ICP quick and precise registration is achieved on original graphics point sets.Experimental results show that the proposed algorithm performs better than other ICP algorithms in speed and precision of registration and has certain practical values. |
| Keywords: | Map point sets ICP algorithm Multi-scale |
| 本文献已被 CNKI 万方数据 等数据库收录！ | |
