电子学报 ›› 2017, Vol. 45 ›› Issue (9): 2156-2161.DOI: 10.3969/j.issn.0372-2112.2017.09.015

• 学术论文 • 上一篇    下一篇

基于局部二进制模式和图变换的快速匹配算法

赵小强, 岳宗达   

  1. 兰州理工大学电气工程与信息工程学院, 甘肃兰州 730050
  • 收稿日期:2016-01-06 修回日期:2016-03-23 出版日期:2017-09-25
    • 作者简介:
    • 赵小强,男,1969年生于陕西岐山.博士.现为兰州理工大学教授、博士生导师.主要研究方向为图像处理、数据挖掘、生产调度.E-mail:xqzhao@lut.cn;岳宗达,男,1990年生于山东阳谷.现为兰州理工大学电气工程与信息工程学院硕士研究生.主要研究方向为图像处理.E-mail:18793139585@163.com
    • 基金资助:
    • 国家自然科学基金 (No.51265032,No.61263003)

A Fast Matching Algorithm Based on Local Binary Patterns and Graph Transformation

ZHAO Xiao-qiang, YUE Zong-da   

  1. College of Electrical and Information Engineering, Lanzhou University of Technology, Lanzhou, Gansu 730050, China
  • Received:2016-01-06 Revised:2016-03-23 Online:2017-09-25 Published:2017-09-25
    • Supported by:
    • National Natural Science Foundation of China (No.51265032, No.61263003)

摘要: 针对图像匹配在图像拼接、目标识别等领域的应用中尺度不变特征变换(Scale Invariant Feature Transform,SIFT)算法计算复杂度高、实时性较差的问题,提出了一种基于局部二进制模式(Local Binary Patterns,LBP)和图变换(Graph Transformation Matching,GTM)的匹配算法.首先采用SIFT特征检测提取特征点并以特征点为中心取13×13的图像块作为特征区域;然后用本文提出的局部旋转不变二进制模式(Local Rotation Invariant Binary Patterns,LRIBP)描述子对特征区域进行描述产生29维的特征描述向量,降低了描述子的复杂度,并以欧氏距离为度量准则进行初始匹配;最后采用图变换匹配算法剔除误匹配点,从而提高算法的运算速率和匹配精度.仿真结果表明,本文所提算法不仅具有较高的精度和较强的鲁棒性,并且减少了算法的运算量,提高了算法的实时性.

关键词: 图像匹配, 尺度不变特征变换, 局部二进制模式, 图变换匹配

Abstract: Aiming at problems of large computation and poor real-time of scale invariant feature transform (SIFT) algorithm for image matching application in image matching,object recognition and other fields,a matching algorithm based on local binary patterns (LBP) and graph transformation matching (GTM) is proposed.Firstly,SIFT is used to extract initial feature points.13×13 pixel blocks around the feature points are used as the feature regions.Secondly,in order to reduce complexity of descriptors,the local rotation invariant binary patterns (LRIBP) descriptor is used to produce feature vectors of 29 dimensions for a feature region.Euclidean distance is adopted as measure criterion of the descriptors to fulfil initial match.Finally,GTM is adopted to eliminate mismatching points.Simulation results show that the proposed algorithm not only improves accuracy and robustness and real-time,but also reduces the amount of calculation.

Key words: image matching, scale invariant feature transform, local binary patterns, graph transformation matching

中图分类号: