当前位置: 首页 > 文章 > 面向单视频超分辨率重建的改进的SIFT配准算法 河北科技师范学院学报 2015,29 (3) 74-79
Position: Home > Articles > The SIFT Registration Algorithm of Super Resolution Reconstruction of Single Video Journal of Hebei Normal University of Science & Technology 2015,29 (3) 74-79

面向单视频超分辨率重建的改进的SIFT配准算法

作  者:
孙文华
单  位:
南昌工程学院信息工程学院
关键词:
特征点检测;特征匹配;SIFT配准算法;超分辨率重建
摘  要:
针对单视频超分辨率重建传统的基于SIFT特征点检测的配准,在特征匹配上占用时间很大,难于满足实时性的要求,而用于计算128维特征描述子的时间最长这一缺点,提出了一种改进的SIFT配准算法,该算法的原理为通过放弃对128维特征描述子的计算,采用特征点之间的灰度相关系数来进行初始匹配,以期大大减少计算时间,提高特征匹配速度。实验结果证明,该算法在配准结果、时间消耗上明显优于传统配置算法。
译  名:
The SIFT Registration Algorithm of Super Resolution Reconstruction of Single Video
作  者:
SUN Wen-hua;School of Information Engineering,Nanchang Institute of Technology;
关键词:
feature point detection;;feature matching;;SIFT algorithm;;super resolution reconstruction
摘  要:
In view of the traditional single video super-resolution reconstruction based on SIFT feature point detecting registration,we found that holding time on the feature matching is too long to meet the requirements of real-time,and the time calculating 128 d features is the longest,thus,we proposed an improved SIFT registration algorithm,which adapted the gray correlation coefficient between the feature points to do the initial matching instead of using the calculation of 128 d feature descriptor,in order to greatly reduce the computing time and increase the speed of feature matching. The experimental results showed that the algorithm is superior to the traditional configuration algorithm on the registration results and time consumption.

相似文章

计量
文章访问数: 6
HTML全文浏览量: 0
PDF下载量: 0

所属期刊

推荐期刊