计算机科学 ›› 2013, Vol. 40 ›› Issue (7): 270-272.

• 图形图像与模式识别 • 上一篇    下一篇

一种改进的SIFT算法

吴建,马跃   

  1. 重庆邮电大学 重庆400065;重庆大学软件学院 重庆401331
  • 出版日期:2018-11-16 发布日期:2018-11-16
  • 基金资助:
    本文受国家自然科学基金(610-71118),重庆邮电大学自然科学基金(A2009-62)资助

Improved SIFT Algorithm

WU Jian and MA Yue   

  • Online:2018-11-16 Published:2018-11-16

摘要: 特征提取是数字图像处理和计算机视觉中的一项重要技术,而利用特征描述算子来构造图像特征点是图像特征提取及配准中的一个关键步骤。SIFT特征点检测算子具有平移、旋转及缩放不变性,在图像配准中应用很广泛。针对基于SIFT特征的64维描述算子的不足进行了改进。通过仿真实验证明,改进后的算法比原算法精度更高,且时间复杂度有所降低。

关键词: 特征提取,特征点,SIFT,图像配准 中图法分类号TP3.05文献标识码A

Abstract: Feature extraction is an important technology in digital image processing and computer vision.And making use of feature descriptor to construct the image feature point is a crucial step in the image feature extraction and image registration.SIFT feature point detection operator has the advantages of translation,rotation and scaling invariance.So,it is widely used in image registration.We mainly improved the 64dimensional description operator based on the SIFT characteristics.The simulation results prove that the improved algorithm has higher accuracy than the original algorithm,and the time complexity is reduced.

Key words: Feature extraction,Feature point,SIFT,Image registration

[1] Lowe D G.Object recognition from local scale-invariant features[C]∥Proceedings of the International Conference on Computer Vision.Corfu,1999:1150-1157
[2] Lowe D G.Distinctive Image Features from Scale-Invariant Key-points [J].International Journal of Computer Vision,2004,0(2):91-110
[3] 杨恒,王庆.一种新的局部不变特征检测和描述算法[J].计算机学报,2010,3(5):935-944
[4] Tang C M,Dong Y,Su X H.Automatic Registration based on Improved SIFT for Medical Microscopic Sequence Images[J].Second International Symposium on Intelligent Information Technology Application,2008,1:580-583
[5] 李志刚.边界重叠图像的一种快速拼接算法[J].计算机工程,2000,6(5):37-38
[6] 魏志强.基于点特征的序列图像匹配方法研究[J].中国图象图形学报,2009,4(3):525-530
[7] 庄志国,孙惠军,董继扬,等.基于特征点检测的图像匹配算法及其在图像拼接中的应用[J].厦门大学学报:自然科学版,2007,6(4):501-505
[8] 胡海青,谭建龙,朱亚涛,等.改进SIFT算法在文字图像匹配中的应用[J].计算机工程,2013,9(1):239-243

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!