首页 > 论文 > 激光与光电子学进展 > 55卷 > 9期(pp:91007--1)

一种改进的KAZE特征检测描述算法

An Improved KAZE Feature Detection and Description Algorithm

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

KAZE特征检测与描述算法在图像匹配方面具有较好的性能。然而, KAZE算法中Perona-Malik(P-M)模型的解不具有唯一性, 而且图像弱边缘在尺度空间中易被平滑。为此, 提出一种改进的KAZE特征检测描述算法(CKAZE)。首先, 基于KAZE原理与能量泛函构建自适应扩散滤波函数; 然后, 研究解的唯一性及图像滤波过程中的边缘保持能力; 最后, 提出CKAZE算法, 利用Mikolajczyk标准数据库图像进行特征匹配实验, 对其性能进行验证。结果表明, 对高斯模糊、光照、旋转缩放、视觉变换而言, CKAZE算法的特征匹配正确率分别较KAZE算法高4.555%、2.138%、0.656%、1.981%, 特征检测和描述的精度提高。

Abstract

For image matching, the KAZE feature detection and description algorithm has demonstrated a number of advantages. However, the solution of Perona-Malik (P-M) model adopted by KAZE is not unique, and the weak edges of image are prone to be smoothed in scale spaces by nonlinear diffusion filter function when the feature points are detected. To overcome these problems, an improved KAZE feature detection and description algorithm for image matching (CKAZE) is proposed. Firstly, an adaptive diffusion filter is built based on the principle of KAZE and energy functional. Then, the solution uniqueness and the edge preserving capacity of the proposed adaptive diffusion filter function are studied during filtering process. Finally, the CKAZE is constructed and its performance is validated through image matching experiments on Mikolajczyk benchmark image dataset. The results demonstrate that the correct rates of feature matching through CKAZE is 4.555%, 2.138%, 0.656% and 1.981% higher, respectively, than those by KAZE for Gauss blurring, illumination, rotation zoom and visual transformation, which indicate that the accuracy of feature detection and description is improved by CKAZE.

Newport宣传-MKS新实验室计划
补充资料

中图分类号:TP391.4

DOI:10.3788/lop55.091007

所属栏目:图像处理

基金项目:安徽省高校自然科学研究重大项目(KJ2017ZD42)、安徽省自然科学基金(1808085ME125)、安徽建筑大学博士启动基金(2015QD04)、安徽省高校自然科学研究项目(KJ2018A0519)

收稿日期:2018-03-19

修改稿日期:2018-04-09

网络出版日期:2018-04-23

作者单位    点击查看

汪方斌:安徽建筑大学机械与电气工程学院, 安徽 合肥 230601安徽建筑大学建筑机械故障诊断与预警重点实验室, 安徽 合肥 230601
储朱涛:安徽建筑大学机械与电气工程学院, 安徽 合肥 230601安徽建筑大学建筑机械故障诊断与预警重点实验室, 安徽 合肥 230601
朱达荣:安徽建筑大学机械与电气工程学院, 安徽 合肥 230601安徽建筑大学建筑机械故障诊断与预警重点实验室, 安徽 合肥 230601
刘涛:安徽建筑大学机械与电气工程学院, 安徽 合肥 230601安徽建筑大学建筑机械故障诊断与预警重点实验室, 安徽 合肥 230601
徐德军:安徽建筑大学机械与电气工程学院, 安徽 合肥 230601安徽建筑大学建筑机械故障诊断与预警重点实验室, 安徽 合肥 230601
许露:安徽省建筑科学研究设计院, 安徽 合肥 230001

联系人作者:汪方斌(wangfb@ahjzu.edu.cn); 储朱涛(zhutaochu@ahjzu.edu.cn);

【1】Patel M I, Thakar V K, Shah S K. Image registration of satellite images with varying illumination level using HOG descriptor based SURF[J]. Procedia Computer Science, 2016, 93: 382-388.

【2】Hsu W Y. A hybrid approach for brain image registration with local constraints[J]. Integrated Computer-Aided Engineering, 2016, 24(1): 73-85.

【3】Chen B, Sun T Q, Liu A X. Influence of speckle noise on image registration based on feature point matching[J]. Laser & Optoelectronics Progress, 2017, 54(12): 121103.
陈波, 孙天齐, 刘爱新. 散斑噪声对基于特征点匹配的图像配准的影响[J]. 激光与光电子学进展, 2017, 54(12): 121103.

【4】Jin J J, Lu W L, Guo X T, et al. Position registration method of simultaneous phase-shifting interferograms based on SURF and RANSAC algorithms[J]. Acta Optica Sinica, 2017, 37(10): 1012002.
靳京京, 卢文龙, 郭小庭, 等. 基于SURF和RANSAC算法的同步相移干涉图位置配准方法[J]. 光学学报, 2017, 37(10): 1012002.

【5】Seregni M, Paganelli C, Summers P, et al. A hybrid image registration and matching framework for real-time motion tracking in MRI-guided radiotherapy[J]. IEEE Transactions on Biomedical Engineering, 2018, 65(1): 131-139.

【6】Ma J Y, Zhou H B, Zhao J, et al. Robust feature matching for remote sensing image registration via locally linear transforming[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(12): 6469-6481.

【7】Chen J H, Guo W S. Method of panoramic image stitching for theodolite-camera system[J]. Laser & Optoelectronics Progress, 2016, 53(5): 051001.
陈继华, 郭文松. 基于摄影经纬仪的全景图像拼接方法[J]. 激光与光电子学进展, 2016, 53(5): 051001.

【8】Weng R L, Lu J, Hu J L, et al. Robust point set matching for partial face recognition[J]. IEEE Transactions on Image Processing, 2016, 25(3): 1163-76.

【9】Yang X C, Qin X B, Wang J, et al. Building faade recognition using oblique aerial images[J]. Remote Sensing, 2015, 7(8): 10562-10588.

【10】Wang G, Sun X L, Shang Y, et al. A robust template matching algorithm based on best-buddies similarity[J]. Acta Optica Sinica, 2017, 37(3): 0315003.
王刚, 孙晓亮, 尚洋, 等. 一种基于最佳相似点对的稳健模板匹配算法[J]. 光学学报, 2017, 37(3): 0315003.

【11】Zeng D, Fang R, Ge S M, et al. Geometry-based global alignment for GSMS remote sensing images[J]. Remote Sensing, 2017, 9(6): 587.

【12】Lowe D G. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision, 2004, 60(2): 91-110.

【13】Bay H, Ess A, Tuytelaars T, et al. Speeded-up robust features (SURF)[J]. Computer Vision and Image Understanding, 2008, 110(3): 346-359.

【14】Alcantarilla P F, Bartoli A, Davison A J. KAZE features[C]∥European Conference on Computer Vision, 2012: 214-227.

【15】Mukherjee P, Lall B. Saliency and KAZE features assisted object segmentation[J]. Image and Vision Computing, 2017, 61: 82-97.

【16】Li D, Zhang M, Tong Z B. Transmission line insulators detection based on KAZE algorithm[J]. Advanced Materials Research, 2014, 936: 2184-2189.

【17】Nabiyev V V, Ylmaz S, Günay A, et al. Shredded banknotes reconstruction using AKAZE points[J]. Forensic Science International, 2017, 278: 280-295.

【18】Liu Y, Lan C, Li C, et al. S-AKAZE: an effective point-based method for image matching[J]. Optik, 2016, 127(14): 5670-5681.

【19】Perona P, Malik J. Scale-space and edge detection using anisotropic diffusion[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1990, 12(7): 629-639.

【20】Xiao D, Huang Y Q. Improved anisotropic diffusion image denoising algorithm[J]. Process Automation Instrumentation, 2017, 38(7): 1-3.
肖丹, 黄玉清. 改进的各向异性扩散图像去噪算法[J]. 自动化仪表, 2017, 38(7): 1-3.

【21】Xu X Z, Zhao W C, Xu F Q, et al. Improved speckle reducing anisotropic diffusion for ultrasound image filtering[J]. Optics and Precision Engineering, 2017, 25(6): 1662-1668.
许贤泽, 赵文成, 徐逢秋, 等. 改进的各向异性扩散方程的超声图像滤波方法[J]. 光学 精密工程, 2017, 25(6): 1662-1668.

【22】Weickert J. Efficient image segmentation using partial differential equations and morphology[J]. Pattern Recognition, 2001, 34(9): 1813-1824.

【23】Yu F, Zhan X Q, Liu M. Stereo vision tracking and measurement system based on KAZE feature[J]. Computer Measurement & Control, 2016, 24(2): 30-33.
于菲, 战兴群, 刘铭. 基于非线性扩散滤波的双目视觉跟踪与测量研究实现[J]. 计算机测量与控制, 2016, 24(2): 30-33.

【24】Weickert J. Applications of nonlinear diffusion in image processing and computer vision[J]. Acta Mathematica Universitatis Comenianae, 2001, 1: 33-50.

【25】You Y L, Kaveh M. Fourth-order partial differential equations for noise removal[J]. IEEE Transactions on Image Processing, 2000, 9(10): 1723-1730.

【26】Liu C J, Zou H L, Qian X. Image processing geometric variational and multiscale method[M]. Beijing: Tsinghua University Press, 2016.
柳婵娟, 邹海林, 钱旭. 图像处理的几何变分与多尺度方法[M]. 北京: 清华大学出版社, 2016.

【27】Mikolajczyk K, Schmid C. A performance evaluation of local descriptors[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 27(10): 1615-1630.

引用该论文

Wang Fangbin,Chu Zhutao,Zhu Darong,Liu Tao,Xu Dejun,Xu Lu. An Improved KAZE Feature Detection and Description Algorithm[J]. Laser & Optoelectronics Progress, 2018, 55(9): 091007

汪方斌,储朱涛,朱达荣,刘涛,徐德军,许露. 一种改进的KAZE特征检测描述算法[J]. 激光与光电子学进展, 2018, 55(9): 091007

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF