作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2012, Vol. 38 ›› Issue (01): 220-221,224. doi: 10.3969/j.issn.1000-3428.2012.01.071

• 图形图像处理 • 上一篇    下一篇

一种基于点的快速图像配准算法

郑雪梅1a,范 勇1a,石琦凯2,柴立群2,王俊波1b   

  1. (1. 西南科技大学 a. 计算机科学与技术学院;b. 理学院,四川 绵阳 621010; 2. 中国工程物理研究院激光聚变研究中心,四川 绵阳 621010)
  • 收稿日期:2011-05-30 出版日期:2012-01-05 发布日期:2012-01-05
  • 作者简介:郑雪梅(1985-),女,硕士研究生,主研方向:图像处理,图像配准,机器视觉;范 勇,副教授、博士;石琦凯、柴立群,副研究员;王俊波,教授、博士
  • 基金资助:
    国家自然科学基金-中国工程物理研究院联合基金资助项目(10676029, 10776028)

Fast Image Registration Algorithm Based on Points

ZHENG Xue-mei 1a, FAN Yong 1a, SHI Qi-kai 2, CHAI Li-qun 2, WANG Jun-bo 1b   

  1. (1a. School of Computer Science and Technology; 1b. School of Sciences, Southwest University of Science and Technology, Mianyang 621010, China; 2. Research Center of Laser Fusion, China Academy of Engineering Physics, Mianyang 621010, China)
  • Received:2011-05-30 Online:2012-01-05 Published:2012-01-05

摘要: 提出一种基于点的快速配准算法。该算法在特征点提取时采用加速分段测试特征算法,通过对比度纹理直方图对特征点进行描述。为加快算法的匹配速度,选用最优节点优先算法进行查找。同时为提高匹配算法的鲁棒性,利用随机抽样一致性算法去除误匹配点对。实验结果表明,与经典的SIFT算法和SURF算法相比,该算法在保持算法稳定性能的同时,可有效提高匹配速度。

关键词: 加速分段测试特征, 对比度纹理直方图, K-D树, 最优节点优先, 随机抽样一致性

Abstract: A fast registration algorithm based on points is presented. Features from Accelerated Segment Test(FAST) is adopted to extract feature points. Contrast Context Histogram(CCH) is exploited to describe feature points. Best Bin First(BBF) searching algorithm is used to accelerate matching. To improve the robustness of the algorithm, Random Sample Consensus(RANSAC) is utilized to remove the mismatched points. Experimental results show that, compared with other registration algorithm such as SIFT and SURF, the proposed algorithm increases the speed observably while keeping the stability of the algorithm.

Key words: Features from Accelerated Segment Test(FAST), Contrast Context Histogram(CCH), K-D tree, Best Bin First(BBF), Random Sample Consensus(RANSAC)

中图分类号: