| 一种新的快速图像匹配算法 |
| |
| 引用本文： | 阳方林,杨风暴,韦全芳,韩焱.一种新的快速图像匹配算法[J].计算机工程与应用,2005,41(5):51-52. |
| |
| 作者姓名： | 阳方林 杨风暴 韦全芳 韩焱 |
| |
| 作者单位： | 中北大学电子信息工程系,太原,030051;中北大学电子信息工程系,太原,030051;中北大学电子信息工程系,太原,030051;中北大学电子信息工程系,太原,030051 |
| |
| 基金项目： | 国家自然科学基金项目资助(编号:60372073),山西高校科技研究开发项目资助(编号:200325) |
| |
| 摘 要： | 目前图像匹配的方法主要有两类:一种是基于灰度的图像匹配,这种方法匹配概率高,但速度较慢;另一类是基于特征的图像匹配,这种方法匹配速度快,但匹配概率不高。文章正是针对基于灰度的图像匹配速度较慢的弊端,首先给出了“图像迹”的定义,然后通过计算图像的迹差,求累计误差的最小值进行图像匹配。通过试验验证,这种方法在保持高的匹配概率的前提下,大大地提高了匹配速率,这种方法只有和差运算,可用DSP等硬件做到实时匹配。
|
| 关 键 词： | 图像匹配 图像迹 匹配概率 累计误差 |
| 文章编号： | 1002-8331-(2005)05-0051-02 |
A Novel and Accelerated Algorithm for Image Match |
| |
| Yang Fanglin,Yang Fengbao,Wei Quanfang,Han Yan.A Novel and Accelerated Algorithm for Image Match[J].Computer Engineering and Applications,2005,41(5):51-52. |
| |
| Authors: | Yang Fanglin Yang Fengbao Wei Quanfang Han Yan |
| |
| Abstract: | Now,there are two methods for image match.One matching method is based image gray.The advantage of the method is higher matching probability,and the disadvantage is slower matching velocity.The other one is based image feature.This method is faster,but lower matching probability.To overcome the disadvantage of image match based image gray,the paper first gives the definition of "image trace",then by computing the trace difference of images and geting the least value of accumulative errors,it finishes the image matching.The tests show that the method not only keeps higher matching probability and greater improves the matching velocity,but has very good anti-noise properties.Because the method has only add and minus operation,it can be easily implemented by DSP hardware etc. |
| |
| Keywords: | image match image trace matching probability accumulative errors |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
