| 基于Hough变换及SIFT特征提取的图像匹配方法 |
| |
| 引用本文： | 苗彩霞,;纪筱鹏.基于Hough变换及SIFT特征提取的图像匹配方法[J].山东科技大学学报(自然科学版),2014(4):105-110. |
| |
| 作者姓名： | 苗彩霞 ;纪筱鹏 |
| |
| 作者单位： | [1]青岛职业技术学院信息学院,山东青岛266555; [2]中国海洋大学信息科学与工程学院,山东青岛266110 |
| |
| 基金项目： | 国家自然科学基金项目（61202208） |
| |
| 摘 要： | 针对现有SIFT算法时间复杂度较高的问题,提出一种基于Hough变换及SIFT特征提取的图像匹配方法。首先,用Hough变换算法检测建筑物区域,以缩小检测与匹配的范围;然后,用SIFT算法在给定区域进行特征点检测与匹配;最后,提出一种两级排除错误匹配的方法,该算法对建筑物序列图像匹配具有光照强度、平移、旋转不变性。实验结果表明,该方法的匹配准确率至少高出比较方法9%。
|
| 关 键 词： | Hough变换 尺度不变特征转换 图像匹配 误匹配去除 数字城市 |
An Image Matching Algorithm Based on Hough Transform and SIFT |
| |
| Institution: | Miao Caixia ,Ji Xiaopeng ( 1. College of Informalion, Qingdao Vocational and Technical College, Qingdao, Shandong 266555, China ; 2. College of Information Science and Engineering, Ocean University of China, Qingdao, Shandong 266110 ,China) |
| |
| Abstract: | Image matching algorithm based on SIFT demands large computation and the amount of matching points is often not reliable. This paper proposes a new image matching method based on Hough transform and SIFT. The building detection algorithm based on Hough transform was presented to confine the area of detecting and matching. Then SIFT algorithm was performed to detect the feature points and match them. Lastly, a new method to avoid mismatching was summarized. This algorithm is invariant to image scale, rotation and change in illumination. The experimental results demonstrate this method can greatly improve computation efficiency and the matching veracity. |
| |
| Keywords: | Hough transform scale invariant feature transform （SIFT） image matching mismatching removing digital city |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
