Image feature point matching algorithm based on center surround filter detection

doi:10.11772/j.issn.1001-9081.2017.12.3547

Abstract

Abstract: Aiming at the problems of poor stability and accuracy of feature point detection in traditional image matching algorithms, a new image feature point matching algorithm based on Scale-invariant Center surround Filter Detection (SCFD) was proposed. Firstly, a multi-scale space was constructed, a center surround filter was used to detect feature points of a image at different scales, Harris method and sub-pixel interpolation were applied to acquire the stable feature points. Secondly, Oriented fast and Rotated Binary Robust Independent Elementary Feature (BRIEF) (ORB) algorithm was combined to confirm the main direction of feature points and construct the description operator of feature points. Finally, Hamming distance was used to complete the matching, Least Median Squares (LMeds) theorem and Maximum Likelihood (ML) estimation were used to eliminate wrong matching points. The experimental results show that, the matching precision of the proposed algorithm is up to 96.6%, which is 2 times of that of the ORB algorithm when the scale changes. The running time of the proposed algorithm is 19.8% of that of Scale Invariant Feature Transform (SIFT) and 28.3% of that of Speed-Up Robust Feature (SURF). The proposed algorithm can effectively improve the stability and accuracy of feature point detection, and has better matching effects under the circumstances of different angle of view, scale scaling, rotation change and brightness variation.

Key words: feature point matching, scale invariance, feature point detection, Oriented fast and Rotated Binary Robust Independent Elementary Feature (BRIEF)(ORB), Least Median Squares (LMedS) theorem

摘要： 针对传统图像匹配算法特征点检测稳定性和准确性差的问题，提出一种尺度不变性的基于中心环绕滤波器检测（SCFD）的图像特征点匹配算法。首先，构建多尺度空间，利用中心环绕滤波器检测图像在不同尺度下的特征点，采用Harris方法和亚像素插值获得稳定的特征点；其次，联合快速定向旋转二进制稳健基元独立特征（BRIEF）（ORB）算法确定特征点的主方向，构建特征点描述算子；最后，采用汉明距离完成匹配，通过最小平方中值（LMedS）定理和最大似然（ML）估计剔除误匹配点。实验结果表明，在尺度变化时，所提算法的匹配精度达到96.6%，是ORB算法的2倍；其运行时间是尺度不变特征变换（SIFT）的19.8%，加速鲁棒性特征（SURF）的28.3%。所提算法能够有效提高特征点检测的稳定性和准确性，在视角、尺度缩放、旋转、亮度等变化的情况下具有较好的匹配效果。

关键词: 特征点匹配, 尺度不变性, 特征点检测, 快速定向旋转二进制稳健基元独立特征, 最小平方中值定理

CLC Number:

TP391.4

SUN Zengyou, DUAN Yushuai, LI Ya. Image feature point matching algorithm based on center surround filter detection[J]. Journal of Computer Applications, 2017, 37(12): 3547-3553.

孙增友, 段玉帅, 李亚. 基于中心环绕滤波器检测的图像特征点匹配算法[J]. 计算机应用, 2017, 37(12): 3547-3553.

References

[1] SINHA S N, FRAHM J M, POLLEFEYS M. Feature tracking and matching in video using programmable graphics hardware[J]. Machine Vision and Applications, 2011, 22(1):207-217.
[2] ZHOU W, LI H, HONG R, et al. BSIFT:towards data-independent codebook for large scale image search[J]. IEEE Transactions on Image Processing, 2015, 24(3):967-979.
[3] CUMMINS M, NEWMAN P. Appearance-only SLAM at large scale with FAB-MAP 2.0[J]. International Journal of Robotics Research, 2011, 30(9):1100-1123.
[4] ALONSO I P, LLORCA D F, GAVILAN M, et al. Accurate global localization using visual odometry and digital maps on urban environments[J]. IEEE Transactions on Intelligent Transportation Systems, 2012, 13(4):1535-1545.
[5] 高晶,吴育峰,吴昆,等.基于角点检测的图像匹配算法[J].仪器仪表学报,2013,34(8):1717-1725.(GAO J, WU Y F, WU K, et al. Image matching method based on corner detection[J]. Chinese Journal of Scientific Instrument, 2013, 34(8):1717-1725.)
[6] 杨晓敏,吴炜,卿粼波,等.图像特征点提取及匹配技术[J].光学精密工程,2009,17(9):2276-2282.(YANG X M, WU W, QING L B, et al. Image feature extraction and matching technology[J]. Optics and Precision Engineering, 2009, 17(9):2276-2282.)
[7] MORAVEC H P. Rover visual obstacle avoidance[C]//Proceedings of the 19817th International Joint Conference on Artificial Intelligence. San Francisco, CA:Morgan Kaufmann, 1981:785-790.
[8] HARRIS C G, STEPHENS M J. A combined corner and edge detector[C]//Proceedings of the 4th Alvey Vision Conference. Manchester, England:[s.n.], 1988:147-151.
[9] LOWE D G. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision, 2004, 60(2):91-110.
[10] 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 & Pattern Recognition. Piscataway, NJ:IEEE 2004:506-513.
[11] BAY H, TUYTELAARS T, VAN GOOL L. SURF:speeded up robust features[C]//Proceedings of the 9th European conference on Computer Vision. Berlin:Springer, 2006:404-417.
[12] AGRAWAL M, KONOLIGE K, BLAS M R. CenSurE:center surround extremas for realtime feature detection and matching[C]//Proceedings of 200810th European Conference on Computer Vision, LNCS 5305. Berlin:Springer, 2008:102-115.
[13] RUBLEE E, RABAUD V, KONOLIGE K, et al. ORB:an efficient alternative to SIFT or SURF[C]//Proceedings of the 2011 IEEE International Conference on Computer Vision. Piscataway, NJ:IEEE, 2011:2564-2571.
[14] 白雪冰,车进,牟晓凯,等.结合快速鲁棒性特征改进ORB的特征点匹配算法[J].计算机应用,2016,36(7):1923-1926.(BAI X B, CHE J, MU X K, et al. Improved feature points matching algorithm based on speed-up robust feature and oriented fast and rotated brief[J]. Journal of Computer Applications, 2016, 36(7):1923-1926.)
[15] 谭晓波.摄像机标定及相关技术研究[D].长沙:国防科学技术大学,2004:30-34.(TAN X B. Study on camera calibration and its correlation technique[D]. Changsha:National University of Defense Technology, 2004:30-34.)
[16] ROUSSEEUW P J. Least median of squares regression[J]. Journal of the American Statistical Association, 1984, 79(388):871-880.
[17] 佘建国,徐仁桐,陈宁.基于ORB和改进RANSAC算法的图像拼接技术[J].江苏科技大学学报(自然科学版),2015,29(2):164-169.(SHE J G, XU R T, CHEN N. Image stitching technology based on ORB and improved RANSAC algorithm[J]. Journal of Jiangsu University of Science and Technology (Nature Science Edition), 2015, 29(2):164-169.)
[18] MIKOLAJCZYK K, SCHMID C. A performance evaluation of local descriptors[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2005, 27(10):1615-1630.