刘立,谭邦,肖建田.尺度不变特征点提取[J].南华大学学报(自然科学版),2009,23(1):85~88.[LIU Li1,2,TAN Bang1,XIAO Jian-tian1.Scale Invariant Feature Points Detection[J].Journal of University of South China(Science and Technology),2009,23(1):85~88.] |
尺度不变特征点提取 |
Scale Invariant Feature Points Detection |
投稿时间:2008-12-08 |
DOI: |
中文关键词: 金字塔匹配 尺度不变 Harris角点 高斯尺度 |
英文关键词:pyramid matching Scale invariant Harris corners Gaussian scale |
基金项目:国家自然科学基金资助项目(60475024);航天技术创新基金资助项目(2006AA09Z203) |
|
摘要点击次数: 808 |
全文下载次数: 969 |
中文摘要: |
结合传统的角点检测算法,提出一种基于高斯金字塔匹配的尺度不变特征点提取算法PBSI,首先建立高斯金字塔,在每层图像中检测Harris角点,根据高斯尺度理论自上而下找到它们在不同层的对应点,所有层上都有的匹配点就是最终的尺度不变特征点.实验结果表明本文算法具有计算简单、抗噪声能力强、稳定性好的优点. |
英文摘要: |
Combined with traditional corner detection algorithm,a scale-invariant feature extraction algorithm is presented based on Gaussian pyramid matching method(PBSI).After building Gaussian pyramid,the Harris corners is detected in each level.In order to find the counterparts of different level a pyramid matching method is taken top-down according to theory of Gaussian scale.The corners can be found in each level is the final feature points.The results show that the algorithm has low computation cost,strong anti-noise capability and good stability. |
查看全文 查看/发表评论 下载PDF阅读器 |
关闭 |
|
|
|