一种CSS-SIFT复合图像配准算法

李培华; 章盛; 刘玉莉; 钱名思

一种CSS-SIFT复合图像配准算法

李培华^{1, 2, 3, 4},
章盛^{1, 2, 3, 4, ,},
刘玉莉^{1, 2, 3, 4},
钱名思^{1, 2, 3, 4}

1.
中航华东光电有限公司，安徽芜湖 241002
2.
安徽省现代显示技术重点实验室，安徽芜湖 241002
3.
国家特种显示工程技术研究中心，安徽芜湖 241002
4.
特种显示国家工程实验室，安徽芜湖 241002

基金项目:

安徽省科技重大专项项目 17030901053

详细信息

作者简介:
李培华（1982-），男，山东潍坊人，硕士研究生，工程师，主要从事微控制器、嵌入式和图形处理方面的研究

通讯作者:
章盛（1989-），男，安徽芜湖人，硕士研究生，主要从事嵌入式、图像配准和人机交互方面的研究。E-mail:18365393973@163.com

中图分类号: TP391
计量
- 文章访问数: 301
- HTML全文浏览量: 150
- PDF下载量: 37
- 被引次数: 0
出版历程
- 收稿日期: 2020-03-15
- 修回日期: 2020-12-27
- 刊出日期: 2021-01-20

CSS-SIFT Composite Image Registration Algorithm

LI Peihua^{1, 2, 3, 4},
ZHANG Sheng^{1, 2, 3, 4
, ,},
LIU Yuli^{1, 2, 3, 4},
QIAN Mingsi^{1, 2, 3, 4}

1.
AVIC Huadong Photoelectric Co., Ltd., Wuhu 241002, China
2.
Key Laboratory of Modern Display Technology, Wuhu 241002, China
3.
National Special Display Engineering Research Center, Wuhu 241002, China
4.
National Engineering Laboratory of Special Display Technology, Wuhu 241002, China

摘要

摘要: 针对SIFT算法的图像配准耗时长的问题，提出一种CSS-SIFT复合图像配准算法。CSS-SIFT算法首先使用CSS算法检测图像特征，然后，使用优化的SIFT算法生成并降维图像特征描述子，最后，使用基于欧式距离和曼哈顿距离的优化双向匹配算法对图像特征进行匹配。仿真实验条件是通过计算机中仿真软件进行仿真实验，统计图像特征数目、匹配数目、正确匹配数目、配准准确率、配准时间与配准时间下降率共6个指标数据，统计结果表明，CSS-SIFT算法在图像配准准确度方面与传统SIFT算法、传统SURF算法、Forstern-SIFT算法、Harris-SIFT算法、Trajkovic-SIFT算法相当，但在图像配准耗时方面分别降低了58.45%、10.68%、14.84%、16.21%与4.63%，为图像配准提供了一种解决方案。
- 尺度不变特征变换算法 /
- 加速稳健特征算法 /
- 曲率尺度空间算法 /
- 图像配准
Abstract: To address the time-consuming problem of image registration in the scale-invariant feature transform(SIFT) algorithm, a curvature scale space (CSS)-SIFT composite image registration algorithm is proposed in this paper. First, the CSS-SIFT algorithm uses the CSS algorithm to extract image features. Image feature descriptors are then generated and reduced by the optimized SIFT algorithm. Finally, an optimized two-way matching algorithm based on Euclidean and Manhattan distances is used for matching.A simulation experiment is conducted using simulation software, and six parameters of index data are employed, including the number of image features, number of matches, number of correct matches, registration accuracy, registration time, and registration time decline rate. Statistical results show that the CSS-SIFT algorithm performs as well as the following algorithms in terms of accuracy of image registration: traditional SIFT, traditional speeded-up robust features, Forstern-SIFT, Harris-SIFT, and Trajkovic-SIFT. In addition, time-consumption of image registration is reduced by 58.45%, 10.68%, 14.84%, 16.21%, and 4.63%, respectively, thus providing an effective solution for image registration.
- SIFT algorithm /
- SURF algorithm /
- CSS algorithm /
- image registration

HTML全文

图 1 CSS-SIFT算法的流程框图

Figure 1. Flow chart of the CSS-SIFT algorithm

下载: 全尺寸图片幻灯片

图 2 实验一待配准图像

Figure 2. Registration images of the experiment one

下载: 全尺寸图片幻灯片

图 3 源图像的特征检测结果图

Figure 3. The results graphs of source image feature detection

下载: 全尺寸图片幻灯片

图 4 右视图像的特征检测结果图

Figure 4. The results graphs of right image feature detection

下载: 全尺寸图片幻灯片

图 5 下视图像的特征检测结果图

Figure 5. The results graphs of down image feature detection

下载: 全尺寸图片幻灯片

图 6 形变图像的特征检测结果图

Figure 6. The results graphs of deformation image feature detection

下载: 全尺寸图片幻灯片

图 7 旋转图像的特征检测结果图

Figure 7. The result graphs of rotated image feature detection

下载: 全尺寸图片幻灯片

图 8 基于传统SIFT算法的图像特征匹配结果图

Figure 8. The result graphs of image feature matching based on the traditional SIFT algorithm

下载: 全尺寸图片幻灯片

图 9 基于传统SURF算法的图像特征匹配结果图

Figure 9. The result graphs of image feature matching based on the traditional SURF algorithm

下载: 全尺寸图片幻灯片

图 10 基于Forstner-SIFT算法的图像特征匹配结果图

Figure 10. The result graphs of image feature matching based on the Forstner-SIFT algorithm

下载: 全尺寸图片幻灯片

图 11 基于Harris-SIFT算法的图像特征匹配结果图

Figure 11. The result graphs of image feature matching based on the Harris-SIFT algorithm

下载: 全尺寸图片幻灯片

图 12 基于Trajkovic-SIFT算法的图像特征匹配结果图

Figure 12. The result graphs of image feature matching based on the Trajkovic-SIFT algorithm

下载: 全尺寸图片幻灯片

图 13 基于CSS-SIFT算法的图像特征匹配结果图

Figure 13. The result graphs of image feature matching based on the CSS-SIFT algorithm

下载: 全尺寸图片幻灯片

图 14 图像配准的对比分析图

Figure 14. Comparison analysis graphs of image registration

T: Tsukuba visible images； K: Kaptein infrared images； M: MR_T1 medical images；
SIFT: Traditional SIFT algorithm； SURF: Traditional SURF algorithm； F-SIFT: Forstern-SIFT algorithm；
H-SIFT: Harris-SIFT algorithm； T-SIFT: Trajkovic-SIFT algorithm； C-SIFT: CSS-SIFT algorithm

下载: 全尺寸图片幻灯片

表 1 不同规模图像集的图像配准时间、准确率比较

Table 1. The comparison of the image registration time, accuracy of different scale image sets

	100		200		300		400
	time/s	accuracy/%	time/s	accuracy /%	time/s	accuracy /%	time/s	accuracy /%
SIFT	735.6	92.28	1502.6	92.10	2305.9	91.89	2985.9	91.96
SURF	335.9	92.52	685.2	92.46	1008.3	92.09	1388.9	92.33
F-SIFT	344.8	93.02	700.3	92.33	1076.3	92.56	1456.8	93.01
H-SIFT	360.7	92.86	730.4	92.54	1082.6	93.03	1480.6	92.48
T-SIFT	318.6	91.89	640.1	92.36	960.7	92.99	1300.8	92.91
C-SIFT	303.7	92.26	618.9	92.37	930.3	92.38	1240.6	92.56

下载: 导出CSV

参考文献(22)

[1]	谢金哲.增强现实中的图像配准方法研究[D].长沙: 国防科学技术大学, 2014. XIE Jinzhe. Research on Image Registration Methods for Augmented Reality[D]. Changsha: National University of Defense Technology, 2014.
[2]	徐鹏.双目视觉的图像配准与拼接及其应用[D].重庆: 重庆邮电大学, 2019. XU Peng. Image Registration And Stitching of Binocular Vision and Its Application[D]. Chongqing: Chongqing University of Posts and Telecommunications, 2019.
[3]	章盛.图像拼接算法的优化及漫游系统的研究[D].芜湖: 安徽工程大学, 2016. ZHANG Sheng. Optimization of Image Stitching Algorithm and Research of Roaming System[D]. Wuhu: Anhui Polytechnic University, 2016.
[4]	Lowe D G. Object recognition from local scale-invariant features[C]// Proceedings of the International Conference on Computer Vision, 1999: 1150-1157.
[5]	Lowe D G. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision, 2004, 60(2): 91- 110. doi: 10.1023/B:VISI.0000029664.99615.94
[6]	Jolliffe I T. Principal Component Analysis[M]. New York: Springer- Verlag New York Inc, 2002.
[7]	Ke Y, Sukthankar R. PCA-SIFT: A more distinctive representation for local image descriptors[C]//Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004: 511-517.
[8]	Bay H, Tuytelaars T, Gool L V. SURF: Speeded up robust features[C]// Proceedings of the European Conference on Computer Vision, 2006: 404-417.
[9]	刘芳, 武桥, 杨淑媛, 等.结构化压缩感知研究进展[J]. 自动化学报, 2013, 39(12): 1980- 1995. LIU Fang, WU Qiao, YANG Shuyuan, et al. Research advances on structured compressive sensing[J]. Acta Automatica Sinica, 2013, 39(12): 1980-1995.
[10]	杨飒, 杨春玲.基于压缩感知与尺度不变特征变换的图像配准算法[J]. 光学学报, 2014, 34(11): 1110001-1-1110001-5. YANG Sa, YANG Chunling. Image registration algorithm based on sparse random projection and scale-invariant feature transform[J]. Acta Optica Sinica, 2014, 34(11): 1110001s-1-1110001-5.
[11]	赵爱罡, 王宏利, 杨小冈, 等.融合几何特征的压缩感知SIFT描述子[J]. 红外与激光过程, 2015, 44(3): 1085-1091. ZHAO Aigang, WANG Hongli, YANG Xiaogang, et al. Compressed sense SIFT descriptor mixed with geometrical feature[J]. Infrared and Laser Engineering, 2015, 44(3): 1085-1091.
[12]	Trajkovic M, Hedley M. Fast corner detection[C]//Image and Vision Computing, 1988: 75-87.
[13]	韩超, 方露, 章盛.一种优化的图像配准算法[J]. 电子测量与仪器仪表, 2017, 31(2): 178-184. HAN Chao, FANG Lu, ZHANG Sheng. An optimized image registration algorithm[J]. Journal of Electronic Measurement and Instrumentation, 2017, 31(2): 178-184.
[14]	Fischler M A, Bolles R C. Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography[J]. Communications of the ACM, 1981, 24(6): 381-395. doi: 10.1145/358669.358692
[15]	胡为, 刘兴雨.基于改进SIFT算法的单目SLAM图像匹配方法[J]. 电光与控制, 2019, 26(5): 7-13. HU Wei, LIU Xingyu. A Monocular SLAM Image Matching Method Based on Improved SIFT Algorithm[J]. Electronics Optics & Control, 2019, 26(5): 7-13.
[16]	Fjortoft R, Lopes A, Marthon P, et al. An optimal multiedge detectors for SAR Image segmentation[J]. IEEE Transactions on Geoscience and Remote Sensing, 1998, 36(3): 793-802. doi: 10.1109/36.673672
[17]	OTSU N. A threshold selection method from gray level histograms[J]. IEEE Transactions on SMC, 1979, 9(1): 62-69.
[18]	迟英鹏, 刘畅.一种适用于SAR图像配准的改进SIFT算法[J]. 中国科学院大学学报, 2019, 36(2): 259-266. CHI Yingpeng, LIU Chang. An improved SIFT algorithm for SAR image registration[J]. Journal of University of Chinese Academy of Sciences, 2019, 36(2): 259-266.
[19]	陆宗骐, 梁诚.用Sobel算子细化边缘[J]. 中国图象图形学报, 2000, 5(6): 516-520. doi: 10.3969/j.issn.1006-8961.2000.06.015 LU Zongqi, LIANG Cheng. Edge thinning based on Sobel Operator[J]. Journal of Image and Graphics, 2000, 5(6): 516-520. doi: 10.3969/j.issn.1006-8961.2000.06.015
[20]	程德强, 李腾腾, 郭昕, 等.改进的SIFT领域投票图像匹配算法[J]. 计算机工程与设计, 2020, 41(1): 162-168. CHENG Deqiang, LI Tenten, GUO Xin, et al. Improved SIFT voting image matching algorithm[J]. Computer Engineering and Design, 2020, 41(1): 162-168.
[21]	Mokhtarian F, Suomela R. Robust image corner detection through curvature scale space[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1998, 20(12): 1376-1381. doi: 10.1109/34.735812
[22]	Canny J. A computational approach to edge detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1986, 8(6): 679-698.