基于学习不变特征变换的兵马俑图像分区匹配

冯筠; 延瑜瑜; 赵妍; 肖方; 刘晓宁

doi:10.3788/OPE.20182607.1774

您当前的位置：

首页 >

文章列表页 >

基于学习不变特征变换的兵马俑图像分区匹配

信息科学 | 更新时间：2020-08-13

- 基于学习不变特征变换的兵马俑图像分区匹配
- A terracotta image partition matching method based on learned invariant feature transform
- 光学精密工程 2018年26卷第7期页码：1774-1783
- 作者机构：
  
  西北大学信息科学与技术学院, 陕西西安 710127
- 作者简介：
  
  [ "冯筠(1972-), 女, 江苏邳州人, 教授, 博士生导师, 1994年于西安电子科技大学获得学士学位, 1997年于西北大学获硕士学位, 2006年于香港城市大学获得博士学位, 主要从事图像、图形处理, 三维重建, 人工智能及模式识别等, 特别是在医学图像影像处理、识别和可视化等方面的研究。E-mail:fengjun@nwu.edu.cn" ]
  [ "刘晓宁(1978-), 女, 陕西省眉县人。西北大学信息科学与技术学院副教授, 硕士生导师, 中国计算机学会会员, 2006年在西北大学获工学博士学位。主要研究方向为图像处理、模式识别与三维可视化技术。E-mail:xnliu@nwu.edu.cn" ]
- 基金信息：
  
  国家自然科学青年基金资助项目(61602380)
- DOI：10.3788/OPE.20182607.1774
  中图分类号： TP391
- 收稿日期：2017-11-27，
  
  录用日期：2018-1-24，
  
  纸质出版日期：2018-07-25
- 稿件说明：
移动端阅览
冯筠, 延瑜瑜, 赵妍, 等. 基于学习不变特征变换的兵马俑图像分区匹配[J]. 光学精密工程, 2018,26(7):1774-1783.

Jun FENG, Yu-yu YAN, Yan ZHAO, et al. A terracotta image partition matching method based on learned invariant feature transform[J]. Optics and precision engineering, 2018, 26(7): 1774-1783.
冯筠, 延瑜瑜, 赵妍, 等. 基于学习不变特征变换的兵马俑图像分区匹配[J]. 光学精密工程, 2018,26(7):1774-1783. DOI： 10.3788/OPE.20182607.1774.

Jun FENG, Yu-yu YAN, Yan ZHAO, et al. A terracotta image partition matching method based on learned invariant feature transform[J]. Optics and precision engineering, 2018, 26(7): 1774-1783. DOI： 10.3788/OPE.20182607.1774.

摘要

针对兵马俑图像三维重建时特征误匹配率较高、效率较低的问题，提出了一种基于两视图兵马俑图像的特征分区匹配方案。该方案在兵马俑图像上，首先利用学习不变特征变换（Learned Invariant Feature Transform，LIFT）方法提取整幅兵马俑图像的特征；接着利用提出的基于先验知识的特征点分布曲线分割方法确定兵马俑头部分割线位置，根据头部分割线将提取的特征分为头部特征和躯干特征；最终采用欧式距离进行分区特征匹配，对于匹配结果集合使用随机抽样一致性算法（Random Sample Consensus，RANSAC）滤除误匹配点集。实验结果表明：在兵马俑图像特征提取及匹配中该方案的正确匹配率可以达到98%，相比于SIFT和SURF方法正确匹配率提高了20%左右，特征点的可重复率提高了10%左右，同时将RANSAC的迭代时间降低了50%，而且在尺度、光照、角度发生变换时具有较好的鲁棒性。因此本文提出的方案能够很好地实现特征点的正确匹配，在兵马俑的三维重建中具有很高的使用价值。

Abstract

A novel feature partition matching scheme for two-view Terracotta warrior images was presented to address the problem of high false matching rate and low feature matching efficiency during 3D reconstruction in this paper. The new scheme was as follows:First

the features of the complete Terracotta warriors image were extracted using the learned invariant feature transform (LIFT) method. Second

the position of the dividing line on the head of the image of the warrior was determined by applying the proposed prior knowledge-based feature point distribution curve

and the extracted features were then divided into head and torso features based on the dividing line. Third

the Euclidean distance was used to perform the regional feature matching

and the random sample consensus (RANSAC) algorithm was subsequently used to filter out the mismatched point set from the matched result set. Experimental results show that in the terracotta image feature extraction and matching

the correct matching rate of the new scheme can reach 98%; the correct matching rate is increased by approximately 20% compared with those of the SIFT and SURF methods

and the repeat rate of the feature points is increased by 10% while the iteration time of RANSAC is decreased by 50%. The new scheme also has better robustness when scale

illumination

and angle are changed in the images. Therefore

the proposed scheme can achieve correct matching of the feature points with sufficient accuracy and has applications in the robust 3D reconstruction of the Terracotta warrior images.

关键词

Keywords

references

张婧. 基于激光扫描数据的遗址场景三维重建技术的研究与应用[D]. 西北大学, 2012. http://cdmd.cnki.com.cn/Article/CDMD-10697-1012444288.htm

ZHANG J. Study and application of three-dimensional reconstruction technology of site scene based on laser scanning data [D]. Northwest University, 2012. (in Chinese)

彭勇, 姚晓伟, 胡勇杰.三维激光扫描技术在石质文物保护中的应用研究[J].城市勘测, 2016(3):97-100.

PENG Y, YAO X W, HU Y J. The application research of 3D laser scanning technology on rockyhistorical relics conservation[J]. Urban Geotechnical Investigation & Surveying, 2016(3):97-100.(in Chinese)

贾棋, 高新凯, 罗钟铉, 等.基于几何关系约束的特征点匹配算法[J].计算机辅助设计与图形学学报, 2015, 27(8):1388-1397.

JIA Q, GAO X K, LUO ZH X, et al.. Feature points matching based on geometric constraints[J].Journal of Computer-Aided Design & Computer Graphics, 2015, 27(8):1388-1397. (in Chinese)

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

BAY H, ESS A, TUYTELAARS T, et al.. Speeded-up robust features[J]. Computer Vision & Image Understanding, 2008, 110(3):404-417.

TOLA E, LEPETIT V, FUA P. A fast local descriptor for dense matching[C]. Computer Vision and Pattern Recognition , 2008. CVPR 2008. IEEE Conference on. IEEE , 2008: 1-8.

RUBLEE E, RABAUD V, KONOLIGE K, et al . . ORB: An efficient alternative to SIFT or SURF[J]. 2011, 58(11): 2564-2571.

MAINALI P, LAFRUIT G, TACK K, et al.. Derivative-based scale invariant image feature detector with error resilience[J]. IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society, 2014, 23(5):2380.

张勇, 耿国华.基于条件数约束的秦俑图像特征点匹配[J].云南大学学报:自然科学版, 2017, 39(4):547-553.

ZHANG Y GENG G H.Feature points matching of terracotta warrior images based on condition number constraint[J]. Journal of Yunnan University(Natural Sciences Edition), 2017, 39(4):547-553. (in Chinese)

冯宏伟, 周亚培, 冯筠, 等.面向多角度视图三维重建的基础矩阵求解[J].光学精密工程, 2016, 24(10s):567-574.

FENG H W, ZHOU Y P, FENG J, et al..Estimation of fundamental matrix from multi-prespective views for 3D reconstruction[J]. Opt. Precision Eng., 2016, 24(10s):567-574.

赵春阳, 赵怀慈.多模态鲁棒的局部特征描述符[J].光学精密工程, 2015, 23(5):1474-1483.

ZHAO CH Y, ZHAO H C. Multimodality robust local feature descriptors[J].Opt. Precision Eng., 2015, 23(5):1474-1483. (in Chinese)

周萌. 秦兵马俑形象在现代卡通动画设计中的应用研究[D]. 西安建筑科技大学, 2013. http://cdmd.cnki.com.cn/Article/CDMD-10703-1014009943.htm

ZHOU M. Study on Application of Qin's Terra Cotta Warriors Image in Modern Animation Design [D]. Xi'an University of Architecture and Technology, 2013. (in Chinese)

YI K M, TRULLS E, LEPETIT V, et al . . LIFT: learned invariant feature transform[J]. 2016: 467-483.

HE K, ZHANG X, REN S, et al . . Deep residual learning for image recognition[C]. Computer Vision and Pattern Recognition. IEEE , 2016: 770-778.

李玉峰, 李广泽, 谷绍湖, 等.基于区域分块与尺度不变特征变换的图像拼接算法[J].光学精密工程, 2016, 24(5):1197-1205.

LI Y F, LI G Z, GU SH H, et al.. Image mosaic algorithm based on area blocking and SIFT[J]. Opt. Precision Eng., 2016, 24(5):1197-1205. (in Chinese)

孟剑明.秦俑百题[M].西安:西北大学出版社, 1997:68-69.

MENG J M. Qin Terracotta Hundred Questions[M]. Xi'an:Northwestern University Press, 1997:68-69.(in Chinese)

屈玉福, 刘子悦, 江云秋, 等.自适应变尺度特征点提取方法[J].光学精密工程, 2017, 25(1):188-197.

QU Y F, LIU Z Y, JIANG Y Q, et al.. Self-adaptative variable-metric feature point extraction method[J].Opt. Precision Eng., 2017, 25(1):188-197.(in Chinese)

李丹, 朱玲玲, 胡迎松.基于最小生成树的多视图特征点快速匹配算法[J].华中科技大学学报(自然科学版), 2017, 45(1):41-45.

LI D, ZHU L L, HU Y S.Fast matching algorithm of multi-view feature points based on minimal spanning tree[J]. Journal of Huazhong University of Science and Technology(Nature Science Edition), 2017, 45(1):41-45.(in Chinese)

郭汉洲, 郭立红, 吕游.结合MSER与HSOG的目标局部特征提取[J].液晶与显示, 2016, 31(11):1070-1078.

GUO H Z, GUO L H, LV Y. Target local feature extraction combined MSER and HSOG[J]. Chinese Journal of Liquid Crystals and Displays, 2016, 31(11):1070-1078.(in Chinese)

COSTANZO A, AMERINI I, CALDELLI R, et al.. Forensic analysis of SIFT keypoint removal and injection[J]. IEEE Transactions on Information Forensics & Security, 2014, 9(9):1450-1464.

HOSSEIN-NEJAD Z, NASRI M. An adaptive image registration method based on SIFT features and RANSAC transform[J]. Computers & Electrical Engineering, 2016.

浏览量

489

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

弱纹理环境下融合点线特征的双目视觉同步定位与建图

针对目标遮挡的自适应特征匹配网络

基于分数阶微分的高光谱图像特征提取与分类

面向高光谱影像场景分类的轻量化深度全局-局部知识蒸馏网络

多分支无锚框网络密集行人检测算法