半导体光电, 2019, 40 (4): 539, 网络出版: 2019-09-20   

融合特征光流与角点特征的图像特征匹配算法研究

Research on Image Feature Matching Algorithm of Fusion Characteristic Optical Flow and Corner Features
作者单位
重庆理工大学 光纤传感与光电检测重庆市重点实验室, 重庆 400054
摘要
图像特征匹配是视觉里程计的重要环节, 针对视觉图像序列特征点匹配中存在的匹配精度低问题, 提出一种融合金字塔特征光流与角点特征的精确快速图像特征匹配算法。算法首先利用ORB(二进制定向简单描述符)算法快速提取图像特征点, 然后融合金字塔Lucas-Kanade特征光流的追踪特性, 使用局部特征窗口计算图像特征点位移矢量。接着针对图像特征的匹配对齐问题以及特征丢失问题, 算法采用K最近邻半径搜索作为特征滤波器移除混淆的匹配, 最后使用RANSAC(Random Sample Consensus)算法剔除冗余误匹配点对, 提高匹配率。通过多组实验数据对比, 该算法的图像特征匹配率可达到98%。对比传统的ORB特征匹配算法, 该算法在实时性和图像特征匹配精度上均有显著提高。
Abstract
Image feature matching is an important part of the visual odometry. To solve the problem of low accuracy of mismatch in the feature point matching of visual image sequences, an accurate and fast algorithm combining the pyramid festure optical flow and corner features is proposed. Firstly, the ORB algorithm is used to extract the image feature points quickly, then the tracking characteristics of the pyramid Lucas-Kanade feature optical flow are fused, and the local feature window is used to calculate the displacement vector of the image feature points. Then, aiming at the matching alignment problem and feature loss problem of image features, the K nearest neighbor radius search is used as the feature filter to remove the confused matching. Finally, the RANSAC (Random Sample Consensus) algorithm is used to eliminate redundant mismatch points to improve the matching rate. By the comparisons of multiple sets of experimental data, the feature matching rate of the proposed algorithm can reach 98%. Compared with the traditional ORB feature matching algorithm, the proposed algorithm realizes a significant improvement in real-time property and accuracy of image feature matching.

赵明富, 陈兵, 宋涛, 曹利波. 融合特征光流与角点特征的图像特征匹配算法研究[J]. 半导体光电, 2019, 40(4): 539. ZHAO Mingfu, CHEN Bing, SONG Tao, CAO Libo. Research on Image Feature Matching Algorithm of Fusion Characteristic Optical Flow and Corner Features[J]. Semiconductor Optoelectronics, 2019, 40(4): 539.

本文已被 1 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!