| 相似性测度函数分析及其在图像匹配中的应用研究 |
| |
| 引用本文： | 宋伟,刘洪亮,纳鹏宇,李明. 相似性测度函数分析及其在图像匹配中的应用研究[J]. 中央民族大学学报(自然科学版), 2014, 0(1): 44-49 |
| |
| 作者姓名： | 宋伟 刘洪亮 纳鹏宇 李明 |
| |
| 作者单位： | [1]中央民族大学信息工程学院,北京100081 [2]中央民族大学理学院,北京100081 [3]北京交通大学电子信息工程学院,北京100044 |
| |
| 基金项目： | 中央民族大学自主科研项目（No.1112KYQN40,No.1112KYZY49）. |
| |
| 摘 要： | 相似性测度函数在计算机视觉领域有着非常重要的意义.本文对距离和相关两类相似性测度的应用频率进行了统计,发现欧氏距离和曼哈顿距离、相关系数应用频率相对较高.从数学角度对三种测度函数的直观意义进行了研究,比较了各测度函数优缺点,探讨了相似性测度函数中各参数变量在图像匹配中具体意义,并基于模板匹配对三种测度函数的实时性进行了仿真比较.实验结果表明曼哈顿距离测度速度最快,欧氏距离次之,相关系数最差.
|
| 关 键 词： | 相似性测度函数 图像匹配 双目视觉 计算机视觉 |
Similarity Measure Function Analysis and Application in Image Matching |
| |
| Affiliation: | SONG Wei, LIU Hong-liang 2 , NA Peng-yu3 , LI ( 1. School of Information Engineering, Minzu University of China, 100081 Ming1 China ; 2. School of Science, Minzu University of China, 100081, China; School of Electronic and Information Engineering, Beijing Jiaotong University, Beijing, 100044, China) |
| |
| Abstract: | Similarity measure function has very important significance in the field of computer vision. In this paper, the application frequency of distance similarity measure and related similarity measure are counted, drawn Euclidean distance, Manhattan distance, and correlation coefficient applied frequently. The intuitive meaning of three measure functions is studied in mathematical sense, and the advantage and disadvantage of them are compared. The specific meaning of similarity measure function＇s variable parameters in image matching is discussed. Based on template matching,the real time of the three measurement function is compared. The resulfs show that Manhattan distance measure is the fastest, Euclidean distance take second place, correlation coefficient is the worst. |
| |
| Keywords: | similarity measure function image matching binocular vision computer vision |
| 本文献已被 维普 等数据库收录！ |
|
