Application of Image Displacement Compensation Based on Feature Block Matching for Medical Injection

doi:10.7517/j.issn.1674-0475.2016.06.541

Abstract

Abstract:

The sequence images captured by the intelligent light inspection machine had the displacement deviation caused by mechanical vibration or tracking asynchronization, which affected the subsequent foreign body segmentation and extraction. To obtain the motion vectors to compensate for the image displacement, the method of feature point detection and block matching is adopted. Firstly, the reference frame was detected by FAST feature points, and used the non-maximal suppression,Hessian matrix,the entropy difference and distance constraint to select stable feature points. According to the location information of the feature points, the matching block were selected, of which the center was located by the feature points. Then, an improved horizontal direction search strategy(HHS) was used to match the current frame to find the best matching block. Finally, the motion parameters were obtained, and the motion vectors were obtained. Experimental results show that the real-time performance of this algorithm is 72 ms, much more faster than the 140 ms algorithm of ORB, and faster than the direct use of the original Hexagonal Search(HS) algorithm for block matching nearly 20%. The accuracy is higher than the feature point matching algorithms. Our algorithm overcomed these problems: 1. the traditional feature point matching could produce a large spatial deviation error matching; 2. the traditional block matching need to preprocess the static background blocks; 3. it is difficult to meet the needs of real time for single feature point matching or block matching. Considering the speed and precision, this algorithm can rapidly compensate the displacement deviation in the image of the bottle, and suppress the interference caused by the deviation of the image displacement.

Key words: machine vision, foreign body detection robot, feature point detection, block matching, search strategy, motion compensation

RUAN Feng, ZHANG Hui, LI Xuanlun, LI Ruoyun. Application of Image Displacement Compensation Based on Feature Block Matching for Medical Injection[J]. Imaging Science and Photochemistry, 2016, 34(6): 541-551.

References

[1] 王耀南, 陈铁健, 贺振东.智能制造装备视觉检测控制方法综述[J]. 控制理论与应用,2015,3(3):273-286. Wang Y N, Chen T J, He Z D. Review on the machine vision measurement and control technology for intelligent manufacturing equipment[J]. Control Theory and Applications, 2015,3(3):273-286.
[2] 吴成中, 王耀南, 贺振东, 张辉. 基于机器视觉的注射液中不溶异物检测方法研究[J]. 仪器仪表学报,2015,7(7):1451-1461. Wu C Z, Wang Y N, He Z D, Zhang H. Research on foreign insoluble particulate detection method for medicinal solution based on machine vision[J]. Chinese Journal of Scientific Instrument, 2015,7(7):1451-1461.
[3] 李博, 王孝通, 杨常青, 金良安. 电子稳像的灰度投影三点局域自适应搜索算法[J].光电工程, 2004,31(9):69-72. Li B, Wang X T, Yang C Q, Jin L A. Three-point locally adaptive searching in gray scale projection algorithm for electronic image stabilization[J]. Opto-electronic Engineering, 2004,31(9):69-72.
[4] 张英涛, 黄剑华, 唐降龙, 刘家锋. 一种基于精简粒子群优化的霍夫变换算法[J]. 天津大学学报,2011,44(2):163-167. Zhang Y T, Huang J H, Tang X L, Liu J F. A hough transform algorithm based on reduced particle swarm optimization[J]. Journal of Tianjin University, 2011,44(2):163-167.
[5] Ko S, Lee S, Jeon S; Kang E. Fast digital image stabilizer based on gray-coded bit-plane matching[J]. IEEE Transactions on Consumer Electronics, 1999,45(3):598-603.
[6] 梁若飞,杨风暴,王毅敏,李大威,冯裴裴. 一种基于块模糊增强的玉米田遥感图像边缘检测方法[J]. 影像科学与光化学,2016,34(2):190-197. Liang R F, Yang F B, Wang Y M, Li D W, Feng P P. An edge detection method for remote sensing images in corn field areas based on blocky fuzzy enhancement[J]. Imaging Science and Photochemistry, 2016,34(2):190-197.
[7] 倪伟, 郭宝龙, 丁贵广. 基于六边形的运动矢量场自适应搜索算法[J]. 计算机工程, 2005,31(13):10-13. Ni W, Guo B L, Ding G G. A hexagon-based motion estimation algorithm using motion vector field adaptive search technique[J]. Computer Engineering, 2005,31(13):10-13.
[8] Lowe D G. Distinctive image features from scale invariant key points[J]. International Journal of Computer Vision, 2004,60(2):91-110.
[9] Herbert B, Andreas E, Tinne T, Luc V G. SURF:Speeded-Up Robust Features[C]. ECCV 2006:Proceedings of the 9th European Conference on Computer Vision. Austria:ECCV, 2006:404-417.
[10] Alahi A, Ortiz R, Vandergheynst P. Freak:fast retina key point[C]. CVPR 2012:Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Colorado Springs:CVPR, 2012:510-517.
[11] Raúl M, Montiel J. M. M, Juan D. T. ORB-SLAM:a versatile and accurate monocular SLAM system[J]. IEEE Transactions on Robotics, 2015,31(5):1147-1163.
[12] Rosten E, Drummond T. Faster and better:a machine learning approach to corner detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010,32(1):105-119.
[13] 张从鹏,魏学光. 基于Harris角点的矩形检测[J]. 光学精密工程, 2014,22(8):2259-2266. Zhang C P, Wei X G. Rectangle detection based on Harris corner[J]. Optics and Precision Engineering, 2014,22(8):2259-2266.
[14] Smith S M, Brady M. SUSAN:a new approach to low level image processing[J]. International Journal of Computer Vision,1997,23(1):45-78.
[15] 杨作廷, 阮萍. 基于图像熵的高动态范围场景的自动曝光算法[J]. 光子学报,2013,42(6):742-746. Yang Z T, Ruan P. Auto-exposure algorithm for scenes with high dynamic range based on image entropy[J].Acta Photonica Sinica,2013,42(6):742-746.
[16] Francesc Aulí-Llinàs. Entropy-based evaluation of context models for wavelet-transformed images[J]. IEEE Transactions on Image Processing, 2015,24(1):57-67.
[17] 刘艳. 基于DS-CDS的改进运动估计算法及实验[J]. 图学学报,2015,36(3):457-461. Liu Y. Improvement of motion estimation algorithm and experiment based on DS-CDS[J]. Journal of Graphics, 2015,36(3):457-461.