用户名: 密码: 验证码:
基于形态学图像技术的群体检测方法研究
详细信息    本馆镜像全文|  推荐本文 |  |   获取CNKI官网全文
摘要
电子产品原材料的加工精度决定了相关产品质量稳定性。由于电子产品集成度越来越高,对原材料的质量提出100%进行检测的要求。其中重点是尺寸与表面质量。传统的图像检测分析方式是针对单独的目标进行处理,图像背景与主体容易区分。作为电子产品的重要原材料之一的“微型陶瓷基材”与其它工业产品不同,它体积很小,检测数量巨大,使用传统的图像检测系统无法满足检测速度的要求。本文针对某陶瓷基材加工企业的产品检测过程进行了研究。主要的研究工作和创新点如下:
     (1)电子基材加工企业生产计划的特点是产品随订单而定,加工产品不固定,产品的规格变化较大,而且每个批次的产量很大,要求检测时间比较短。传统意义上的视觉检测系统不适用于此类产品的检测。在本课题的研究中分析了传统机器视觉检测系统的特点与不足,第一次明确提出了“群体检测”的概念,并对此概念进行了定义与说明,同时与传统检测方式进行了对比。
     (2)由于被检测对象属于微型目标,采样后数据中的背景噪声会严重影响目标的检测。传统检测技术的图像预处理中无论是空域还是频域的滤波都会对边缘的定位产生影响,其中比较明显的是边缘的飘移。论文分析了形态学处理模式的特点,特别是其中结构算子的特征情况,将偏微分方程运用到“群体检测”中。
     (3)传统视觉检测系统中光源的主要功能是强化边缘或目标的表面。陶瓷材料表面具有较强的反光性,当光源选择或设置不当的时候,在数据采集过程中会出现二次的信号污染。研究过程中集中针对不同光源与不同设置方案进行了详细分析对比,提出了针对“群体检测”适宜的解决方式。
     (4)基材生产企业的特点是产量大、生产周期短、产品规格变化频繁。为了使系统具有更强的鲁棒性,本文提出使用面向对象的开发环境,结合通用图像函数库的视觉检测系统开发方案,并通过具体案例的分析说明了系统的架构过程。研究过程中对电子基材产品企业的质量检测方式进行了分析,可以实现在产品规格发生变化的情况下,在原有的检测系统中通过修改软件以保证系统能够重复使用。有效地解决了订单多变的外来加工型企业在应对日益增加的质量检测要求下的技术难题,提升了企业的竞争力。
     (5)研究设计了相应的检测系统,并将检测图像计算的结果与专业的电子显微镜准确测量的结果进行比对,对产生测量误差的原因进行了分析,检测结果符合检测的技术要求。
     通过对检测数据的分析表明,文中提出的基于图像函数库设计的群体视觉检测系统的检测精度能够满足企业的精度与速度的要求,且系统具有较强的通用性。所取得的成果对于微型材料生产企业的检测技术具有一定的理论意义和较大的实用价值。
Electronic products become increasingly integrated and stability of product quality directly depends on the processing accuracy of raw materials, so the qualities of the100%raw materials are demanded to detect (focus on size and surface quality). In traditional detection methods, the image analysis treats a single target and it is easily to distinguish the background from the main body. As one of important raw materials of electronic products, the "micro-ceramic substrate" with the small size and large detection quantities is different from the other industrial products. The general detection system can not meet the requirements of detection speed. In this paper, the cases on some ceramic substrate processing enterprises are researched. The main research work and innovations are as follows:
     (1) Electronic substrate processing enterprises are characterized by products with the order and the products are not fixed. Therefore, in processing there are a greater change in product specifications, the large output of each batch and a relatively short detection time. Traditional visual inspection system does not apply for the detection of these products. In this study, the features and shortcomings of the traditional machine vision inspection system are analyzed. The concept of "group detection" is first clearly stated, defined and compared with conventional testing methods.
     (2) Since the detected object is micro-target, the noise in sampled data seriously affects target detection. Traditional image preprocessing technique in the detection whether the spatial or frequency domain filtering will affect the positioning of the edge, the more obvious case is the drift of edge. In this paper, the characteristics of morphological processing model were analyzed and morphological processing methods were studied, especially the structural characteristics of operators. The partial differential equations for edge positioning were used in "group detection".
     (3) The primary function of light source in traditional visual inspection system is to strengthen the edge or target surface. Because ceramic material has a strong reflective surface, there will be a signal of secondary pollution in the data collection process with poor set or choices in light source. In this study, the different light sources with different combination programs were analyzed and compared in detail and the suitable solution on "group detection" was made.
     (4) Substrate manufacturing enterprises are characterized by yield, short production cycle and the frequent change product specifications. To make the system more robust, the development programs of detection system combined with common image library are used in the object-oriented development environment and the system architecture process is analyzed through the specific cases. In the study, products quality detection in electronic substrates company was explored and in the case of changing product specifications, the software in the original detection system was modified to ensure that the system can be reused. The technical problems of detecting under the increasing quality detection requirements are solved effectively and the competitiveness of enterprises are enhanced.
     (5) The corresponding detection system was researched and the computing results of test image were compared with the accurate measurement of professional electron microscopy. The results basically meet the testing requirements and the causes produced measurement deviations were analyzed.
     The analysis results of test data show that the detection accuracy of group visual inspection system based image library proposed in the paper can meet the requirements of accuracy and detection speed. The achievement obtained has a certain theoretical and great practical value for detection technology in micro-materials enterprises.
引文
[1]李红俊,韩冀皖.数字图像处理技术及其应用[J].计算机测量与控制.2002.10(9):620-622
    [2]周记林,马莉.基于机器视觉的珍珠图像采集及缺陷检测.[J]计算机工程与应用.2008,44(3):210-213
    [3]杨丽凤.机器视觉在基片自动分类机中的应用.[J]机械工程与自动化.2004.6(3) :0018-0019
    [4]张琦.机器视觉系统的原理及现状[J].电子工业专用设备,1999,28(4):20-21.
    [5]宋培华,牛玉渊,高敦岳.在线检测中机器视觉的CCD方案[J].计算机工程,2001,27(11)23-24.
    [6]魏伟波,芮筱亭,杨富锋.基于图像的目标自动跟踪系统.[J]计算机工程与应用.2007,43(27):190-192
    [7]陈虹.汽车车牌的自动检测与识别[J].铁路计算机应用.第18卷第11期:0017-0020
    [8]夏若安,朱理,胡双炎.高精度零件尺寸测量系统[J].科学技术与工程.2005.第5卷第20期:1506-1509
    [9]葛鲁宁,杨建国,王庆霞.图像测量在芯片检测中的应用[J].计算机应用.2009年6月第29卷:0242-0244
    [10]梅辉,邓晓东,孙磊.陶瓷基复合材料无损检测研究进展[J]材料导报.2009年5月(上)第23卷第5期:82-87
    [11]Sun J G,Erdman S,Connolly L.Measurement of delamination size and depth in ceramic matrix composites using pulsed thermal imaging[J]. Ceram Eng Sci Proc,2003,24 (4):201-206
    [12]Abdul2Aziz A, Saury C, Xuan V B, et al. On the material characterization of a composite using micro CT image based finite element modeling [J]. Proc SPIE Int Soc Opt Eng,2006,6176 (5):1
    [13]Schilling P J,Karedla B R, Tatiparthi A K, et al. X-ray computed microtomography of internal damage in fiber reinforced polymer mat rix composites [J]. Comp Sci Techn,2005,65:2071
    [14]Lee H R, Ellingson W A. Characterization of a flat-panel amorphous silicon detector for X-ray computed tomography [J].J X-ray Sci Techn,2001,9 (1):43
    [16]伍济钢,宾鸿赞.薄片零件机器视觉图像亚像素边缘检测.[J].中国机械工程.2009第20卷第3期:0297-0300
    [17]高德威,陈天煌,刘朋.蚁群算法在图像边缘检测上的应用研究[J].计算机与数字工程.2009年第 1期:131-134
    [18]M. Dorigo, V. Maniezzo, A. Colorni. The ant system:Optimization by a colony of cooperating agents[J].IEEE Transactions on Systems, Man, and Cybernetics-part B,1996,26 (1):1213
    [19]Chialvo D R, Millonas M M. How swarms build cognitive maps [J]. The Biology and Technology of Intelligent Autonomous Agent s,1995, (144):439-450
    [20]刘粲,谢小鹏,陆丕清.基于BP神经网络的铁谱磨粒图像识别方法研究[J]润滑与密封.2010年第35卷第4期:0072-0075
    [21]贾方庆,杨天怡,黄勤.基于机器视觉的在线环境监测系统设计[J]计算机工程与设计.2007年第28卷第17期:4303-4306
    [22]黄显川,叶志庭.用于LCD背光模块的LED光源转成线光源设计研究[J].光学与光电技术.2007年4月第5卷:0008-0011
    [23]杨明,白烨,王秋良.面阵CCD摄像机光学镜头参数及选用[J].光电子技术与信息.2005,18(3):0027-2031
    [24]贾金萍,张志东.LED背光源亮度均匀性的研究[J]现代显示(2006)04:0019-0022
    [25]R Tsai. A Versatile Camera Calibration Technique for High Accuracy 3D Machine Vision Metrology Using off-the-shelf tv Cameras and Lenses[R]. Research Report 51342, IBM T.J.Watson Research Center,Yorktown Heights,NY 10598,1985.
    [26]O Faugeras, G Toscani. The Calibration Problemfor Stereo[C]. Proc. Int. Conf. Computer Vision, 1990.1272140.
    [27]H A Martins,J R Birk,R B Kelley. Camera Models Based on Data from Two Calibration Planes [J]. Comp. Vision Graphics and Image Proc.,1981,(17):1732180.
    [28]B Caprile, V Torre. Using Vanishing Points for Camera Calibration[J]. Int. Journal of Computer Vision,1990, (4):1272140
    [29]Z Zhang. A Flexible New Technique for Camera Calibration.Technical Report MSR2TR298271, Microsoft Research, Dec 1998. Available Together with the Software [EB/OL]. http://research. microsoft. com/-zhang/Calib/.
    [30]Daniel E Steven,Margaret M Fleck. Nonparametric Correction of Distortion Application of Computer Vision[C].WACV'96, Proceedings 3rd IEEE Workshop on,1996.2142219.
    [31]吴文琪,孙增圻.机器视觉中的摄像机定标方法综述[J].计算机应用研究.2004.02:0004-0006
    [32]www.ccs-inc.co.jp
    [33]http://www.daheng-image.com/AIIUploadFiles/UploadFile/CQPPT/HALCON 9.0 Introduction.pdf
    [34]http://www.china-vision.net/rencai/qiuzhi/200611/3304.html
    [35]http://www.eefocus.com/article/09-03/68551 s.html
    [36]黄建荣,范彦斌,叶邦彦.基于Halcon的SMT-PCB元件缺陷检测[J].佛山科学技术学院学报(自然科学版)2008.26(2):19-21
    [37]梅迎春,王朝立,廖勤武.基于OpenCV的摄像机标定的应用研究[J].计算机工程与设计,2009,30(16):3856-3858
    [38]王俊骅,张方方,张兰芳.基于OpenCV和Halcon的交通冲突视频自动检测及数据处理[J].同济大学学报(自然科学版),2010,38(2):238-244.
    [39]罗珍茜,薛雷,孙峰杰等.基于HALCON的摄像机标定[J].电视技术.2010,(04):101-102
    [40]冈萨雷斯.数字图像处理(第二版)[M].电子工业出版社.北京.2006
    [41]陆玲,王蕾,桂颖.数字图像处理.[M].中国电力出版社.北京.2007
    [42]张专成,妙全兴,田杰,等.一种基于多尺度噪声检测的图像中值滤波器[J].计算机工程与应用,2008,44(7):60-62.
    [43]Vladimir Crnojevic,Vojin Senk,Zeljen Trpovski.Advanced impulse detection based on pixel-wise MAD[J].IEEE Signal Processing Letter,2004,11 (7):589-592.
    [44]韩晓微,等.一种基于脉冲噪声检测的图像均值滤波方法[J].计算机工程与应用,2004,40(27):102-104.
    [45]李迎春,孙继平.图像脉冲噪声的自适应长距离相关迭代滤波法[J].计算机工程与设计,2007,28(1):103-105.
    [46]黄全品,王绪本.一种高效的基于阈值的图像滤波算法及其实现[J].计算机仿真,2005,22(5):111-114.
    [47]张旭明,徐滨士,董柿运.用于图像处理的自适应中值滤波[J].计算机辅助设计与图像图形学报,2005,17(2):295-299.
    [48]Lin T C,Yu P T.Salt-pepper impulse noise detection and removalusing multiple thresholds for image restoration[J]. Journal of Information Science and Engineering,2006,22:189-198.
    [49]Raymond H Chan,Chung-Wa Ho,Mila Nikolova.Salt and pepper noise removal by median-type noise detectors and detailpreserving regularization [J]. IEEE Trans,2005,14 (10):1479-1485
    [50]MALLAT S. A Wavelet Tour of S i gnal Processing [M].2nd ed. Beijing:China Machine Press, 2003.
    [51]赵瑞珍,刘晓宇,孙民贵.基于稀疏表示的小波去噪[J].中国科学:信息科学,2010年01期:33-40
    [52]高飞,杨平先,孙兴波.基于小波变换与阀值收缩法的图像增强去噪[J].四川理工学院学报(自然科学版),2006(02):08-11
    [53]王欣,王云霄,于晓.基于多小波闽值收缩与子带增强的图像去噪[J]哈尔滨工业大学学报.2008(01):0152-0154
    [54]Gall egos,Funes FJ,Ponomaryov V I. Real time image filtering scheme based on robust estimators in presence of impulsive noise[J]. Real Time Imaging,2004,10 (2):69-80.
    [55]Lukac R, Smolka B,Mart in K. Vector filtering for color imaging[J]. IEEE Signal Processing Magazine,2005,22 (1):74-86.
    [56]Yuan S Q,T an Y H.Im pulse noise removal by a global local noise detect or and adaptive median filter[J]. Signal Processing,2006,86 (8):2123-2128.
    [57]Smolka B, Lukac R, Chydzinski A, et al. Fast adaptive similarity based impulsive noise reduction filter[J].Real Time Imaging,2003,9 (4):261-276.
    [58]QUN C,L IM A,ONG KW. D (k)-index:an adaptive structural summary for graph-structured data[C]//Special Interest Group on Management Of Data Conference. CA:ACM Press,2003:134-144.
    [59]SHASHA D,WANG J T L,GIUGNO R.Algorithmics and app lications of tree and graph searching[C]//Principles of Database Systems. Wisconsin:ACM,2002:39-52.
    [60]YAN X, YU P S, HAN J. Graph indexing:a frequent structure based app roach[C]//Special Interest Group on Management Of Data Conference. Paris:ACM Press,2004:335-346.
    [61]FU K S. A step towards unification of syntactic and statistical pattern recognition [J].Transactions on Pattern Analysis andMachine Intelligence,1986,8 (3):398-404.
    [62]FANG Y, KATZ R H, LAKSHMAN T V. Gigabit rate packet pattern-matching using tcam[C]//International Conference on Network Protocols. Berlin:IEEE,2004:174-183
    [63]蒋爱平,梁舒,马爽.基于形态学和细胞神经网络的边缘检测方法[J].光电工程.2008(10)年10月:076-080
    [64]Gonzalez Rafael, Woods Richard. Digital Image Processing:Second Edition [M]. RUAN Qiu-qi, RUAN Yu-zhi. Beijing:Publishing House of Electronic Industry,2003.
    [65]Koren R, Yitzhaky Y. Automatic selection of edge detector parameters based on spatial and statistical measures [J]. Computer Vision and Image Understanding,2006,102:204-213.
    [66]Staunton R C. Detected edge position evaluation using measured acquisition system parameters [J]. Pattern Recognition Letters,2005,26 (11):1609-1619.
    [67]Serra J. Image Analysis and Mathematical Morphology [M]. New York:Academic Press,1982.
    [68]Serra J. Image Analysis and Mathematical Morphology, Part Ⅱ:Theoretical Advances [M]. New York:Academic Press,1988.
    [69]Serra J. Introduction to mathematical morphology [J]. Computer Vision, Graphics, and Image Process,1986,35:283-305.
    [70]Harrer H, Nossek J A. Discrete-Time Cellular Neural Networks [J]. International Journal of Circuit Theory and Applications,1992,20 (4):453-467.
    [71]ter Brugge M H, Nijhuis J A G, Spaanenburg L. Transformational DT-CNN design from morphological specifications [J]. IEEE Transactions on Circuits and System-I:Fundamental Theory and Applications,1998,45 (9):879-888.
    [72]Jiang Mingyan,Yuan Dongfeng. A Multi-Grade Mean Morphologic Edge Detection[J]. IEEE,ICSP'02 Proceedings,2002,1079-1082.
    [73]James S. J. Lee, R. Haralick,1. shapiro. Morphologic edge detection [J]. IEEE Journal of Robotics and Automation,1987,3 (2):142-156.
    [74]Fu Liu,Jian Deng,Pingyuan Cui. A Method for Lunar Obstacle Detection Based on Multi-scale Morphology Transformation[J]. Proceedings of the 2007 IEEE International Conference on Mechatronics and Automation. August 5-8,2007,Harbin,China.2736-2739.
    [75]Fei Zhang,Yong Lian. Electrocardiogram QRS Detection Using Multiscale Filtering Based on Mathematical Morphology[C].Proceedings of the 29th Annual International Conference of the IEEE EMBS Cite Internationale,Lyon,France August 23-26,2007,3196-3198.
    [76]Malay K. Kundu. Morphologic Edge Detection with Multi-scale Approach[J].IEEE,1998,53-56.
    [77]Antonio Plaza. Dimensionality reduction and classification of hyperspectral image data using sequences of extended morphological profiles[J].IEEE transactions on Geoscience and Remote Sensing,2005,43 (3):466-478.
    [78]康怀祺,史彩成,赵保军.一种基于扩展数学形态学的边缘检测方法[J].光学技术,2006,4(32):635.
    [79]Rafael C. Gonzalez,Richard E.Woods,Steven L. Eddins. Digital Image Processing Using Matlab. Second Edition (M). Beijing:Publishing Hose of Electronics Industry,2005,446.
    [80]Yanghui,Jiwu Zhang. Research on Application of Mathematical Morphology in Edge Detection of Image[J].JOURNAL OF LIAONINGU UIVERSITYNatural Science Edition,2005,32 (1):50-51.
    [81]Zhenhua Li, Yingping Yang. Multi-scale Morphologic Tracking Approach for Edge Dctection[J]. IEEE of Computer Society,2007,86 (10):358-362.
    [82]曾欢,王浩.图像边缘检测算法的性能比较与分析[J].现代电子技术,2006,229:53-55
    [83]丁明跃.蔡超.医学图像处理[M].高等教育出版社.北京.2010
    [84]Caselles V, More J. Introduction to the special issue on partial differential equations and geometry-driven diffusion in image processing and analysis [J].IEEE Transaction on image Processing,1998,7 (3):269-273
    [85]Pernoa P, Malik J. Scale-space and edge detection using anisotropic diffusion. [J]. IEEE Trans. Pattern And. Machine Intelligence,1990,12 (7):629-639.
    [86]Liang P, Y F. Local scale controlled anisotropic diffusion with local noise estimate for image smoothing and edge detection. [M]. Washington, DC, USA, IEEE Computer Society,1998.
    [87]Elder J H, Zucker S W, Local scale control for edge detection and blue estimation, [J]. IEEE Transaction on Pattern Analysis and Machine Intelligence,1998,20 (7):699-716
    [88]You Y L, Xun W Y, Tannenbaum A, et. Behavioral analysis of anisotropic diffusion in image processing, [J]. IEEE Transaction on image Processing,1996,5 (11):1539-1553
    [89]Segall C A, Acton S T. Moorphological anisotropic diffusion, [J]. IEEE International Conference on Image Processing,1997,3:348-351
    [90]Pratt W K. Digital Image Process[M]. Publisher Johm Wiley & Sons,1991:20-23
    [91]蔡超,周成平,丁明跃.一种应用于异质扩散的噪声腐蚀新算子[J].华中科技大学学报(自然科学版),2004,32(11):48-50
    [92]李奇,冯华君,徐之海,等.数字图像清晰度评价函数研究[J].半导体光电,2002,31(6):736-738
    [93]孙双花.视觉测量关键技术及在自动检测中的应用[D].天津:天津大学精密仪器与光电工程学院,2007
    [94]费旭东,王文剑,荆人杰.二值图边界的快速提取算法[J].信号处理,1994,10(3):130-138
    [95]葛澎.一种快速边缘跟踪与提取的新算法[J].微电子学与计算机,2005,22(8):14-17
    [96]刘力双,张铫,卢慧卿,等.图像的快速亚像素边缘检测方法[J].光电子·激光,2005,16(8):993-996

© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号

地址:北京市海淀区学院路29号 邮编:100083

电话:办公室:(+86 10)66554848;文献借阅、咨询服务、科技查新:66554700