红外与激光工程, 2018, 47 (11): 1126006, 网络出版: 2019-01-10  

基于LK和FAST的时间序列图像快速配准算法

Fast registration algorithm of image sequence by time based on LK and FAST
作者单位
1 南京理工大学 自动化学院, 江苏 南京 210094
2 火箭军工程大学 导弹工程学院, 陕西 西安 710025
摘要
LK光流算法是一种精确高效的特征跟踪算法, 能够较大幅度提高图像配准的精度和速度。针对时间序列图像的配准问题, 基于LK光流算法, 通过基于图像金字塔的方式跟踪改进后的FAST特征角点, 采用一种鲁棒的单应矩阵估计算法解算配准参数, 提出了一种基于LK光流和改进FAST特征的实时鲁棒配准算法。通过一组时间序列图像从配准精度和配准速度两个方面对所提出算法的性能进行了验证分析, 平均重投影误差为0.16, 平均处理速度为30 Hz。实验结果表明, 该算法能够提取稳定的FAST角点, 快速准确地跟踪匹配序列图像之间的特征, 较好地解决时间序列图像的实时配准问题。
Abstract
LK optical flow is an accurate and efficient feature tracking method which can be used to improve the performance of the image registration algorithm. For the registration problem of image sequence by time, a real-time and robust registration algorithm combining LK optical flow and improved FAST corners was proposed. The improved FAST corners was tracked by using the LK optical flow based on image pyramid and the registration parameters were calculated by adopting a robust homography estimation algorithm. In the experimental part, a real image sequence by time was used to verify the performance of the proposed algorithm from two aspects: registration accuracy and registration speed. The average re-projection error was 0.16 with the processing speed of 30 Hz. The experimental results show that the proposed algorithm can extract stable FAST corners and match the features between images efficiently and accurately, which solve the real-time registration problem of image sequence by time.
参考文献

[1] Jende P, Nex F, Gerke M, et al. Fully automatic feature-based registration of mobile mapping and aerial nadir images for enabling the adjustment of mobile platform locations in gnss-denied urban environments[J]. The International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, 2017, 42: 317.

[2] Ruppert G C S, Chiachia G, Bergo F P G, et al. Medical image registration based on watershed transform from greyscale marker and multi-scale parameter search[J]. Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2017, 5(2): 138-156.

[3] Chang C H, Chou C N, Chang E Y. CLKN: Cascaded Lucas-Kanade networks for image alignment[C]//The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017.

[4] Modin K, Nachman A, Rondi L. A multiscale theory for image registration and nonlinear inverse problems[J/OL]. arXiv preprint arXiv: 1803.01790, 2018.

[5] Liu S, Jiang J. Remote sensing image registration based on feature points of global edge[C]//Imaging Systems and Techniques (IST), 2017 IEEE International Conference on, IEEE, 2017: 1-6.

[6] Fan J, Wu Y, Li M, et al. SAR image registration using multiscale image patch features with sparse representation[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2017, 10(4): 1483-1493.

[7] Li C, Ma Q, Xiao L, et al. Image splicing detection based on Markov features in QDCT domain[J]. Neurocomputing, 2017, 228: 29-36.

[8] 张鹏强, 余旭初, 韩丽, 等. 基于直线特征匹配的序列图像自动配准[J]. 武汉大学学报(信息科学版), 2007, 32(8): 676-679.

    Zhang Pengqiang, Yu Xuchu, Han Li, et al. Automatic registration of image sequence based on line matching approach[J]. Geomatics and Information Science of Wuhan University, 2007, 32(8): 676-679. (in Chinese)

[9] 王志强, 程红, 杨桄, 等. 全局图像配准的目标快速定位方法[J]. 红外与激光工程, 2015, S1(44): 225-229.

    Wang Zhiqiang, Cheng Hong, Yang Guang, et al. Fast target location method of global image registration[J]. Infrared and Laser Engineering, 2015, 44(S1): 225-229.

[10] 何林阳, 刘晶红, 李刚, 等. 改进BRISK特征的快速图像配准算法[J]. 红外与激光工程, 2014, 43(8): 2722-2727.

    He Linyang, Li Jinghong, Li Gang, et al. Fast image registration approach based on improved BRISK[J]. Infrared and Laser Engineering, 2014, 43(8): 2722-2727. (in Chinese)

[11] 赵辽英, 吕步云, 厉小润, 等. 基于尺度不变特征变换和区域互信息优化的多源遥感图像配准[J]. 物理学报, 2015, 64(12): 190-200.

    Zhao Liaoying, Lv Buyun, Li Xiaorun, et al. Multi-source remote sensing image registration based on scale-invariant feature transform and optimization of regional mutual information[J]. Acta Phys Sin, 2015, 64(12): 190-200. (in Chinese)

[12] 张健, 娄树理, 任建存. 基于时序多帧投影的空间目标检测算法[J]. 激光与红外, 2015, 45(1): 88-93.

    Zhang Jian, Lou Shuli, Ren Jiancun. Space target detection algorithm based on sequential multi-frames projection[J]. Laser & Infrared, 2015, 45(1): 88-93. (in Chinese)

[13] Bouguet J Y. Pyramidal implementation of the affine lucas kanade feature tracker description of the algorithm[J]. Intel Corporation, 2001, 5(1-10): 4.

[14] Rosten E, Drummond T. Machine learning for high-speed corner detection[C]//European Conference on Computer Vision, 2006: 430-443.

[15] 齐乃新, 张胜修, 曹立佳, 等. 一种对错误匹配点鲁棒的多单应矩阵估计方法[J]. 机器人, 2017, 39(5): 608-619.

    Qi Naixin, Zhang Shengxiu, Cao Lijia, et al. An estimation method robust to outliers for multiple homographies[J]. Robot, 2017, 39(5): 608-619. (in Chinese)

[16] Qi Naixin, Zhang Shengxiu, Cao Lijia, et al. Fast and robust homography estimation method with algebraic outlier rejection[J]. IET Image Processing, 2017, 12(4): 552-562.

[17] Szpak Z L, Chojnacki W, Eriksson A, et al. Sampson distance based joint estimation of multiple homographies with uncalibrated cameras[J]. Computer Vision and Image Understanding, 2014, 125: 200-213.

荆滢, 齐乃新, 杨小冈, 卢瑞涛. 基于LK和FAST的时间序列图像快速配准算法[J]. 红外与激光工程, 2018, 47(11): 1126006. Jing Ying, Qi Naixin, Yang Xiaogang, Lu Ruitao. Fast registration algorithm of image sequence by time based on LK and FAST[J]. Infrared and Laser Engineering, 2018, 47(11): 1126006.

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!