基于高分辨率CCD相机的X射线实时图像的处理与分析
|
摘要
随着高分辨率、高性能数字视频相机的成本下降,所在课题组利用高分辨率CCD数字相机和新一代影像增强器组成了高分辨率的X射线实时成像系统,和传统的图像增强器及普通模拟视频CCD相机组成的系统相比,具有空间分辨率高、抗干扰性好等特点。本论文主要工作是开发其检测软件,它具有图像获取及多种处理的功能,而且能够实现对特征参数的计算,并辅助评判人员实现对缺陷级别的判定。
首先,介绍了课题组搭建的一套成像系统的硬件组成。在熟悉CCD数字相机及采集卡以后,编制了图像采集及保存软件,然后根据常规X射线检测的需要和汽车铝轮毂检测项目的实际需要,开发了一整套软件。在图像预处理部分,通过分析图像噪声,分别采用多帧叠加和中值滤波方法,达到了预期的结果。对于对比度不高的问题,采用灰度变换、直方图均衡等方法,对图像进行了增强；针对轮毂图像的特点及常规增强方法的不足,提出了一种基于Curve let变换的增强算法,提高了原图像的对比度,达到了细节区域较好的视觉效果。在图像分割阶段,采用减影技术进行分割,达到很好的分割效果,但是在分割过程中,容易造成缺陷的失真。然后在分析传统边缘提取算法的基础上,提出了基于SUSAN算子的边缘检测和基于目标区域相结合的算法,取得了理想的效果。结合企业对检测的要求及铸件缺陷评判标准,介绍了缺陷的分类、特征表达和表述方法,对周长、面积以及非几何特征进行测量,并设计了一种特征测量的实时显示系统,使缺陷特征能够实时显示出来。另外,还设计了扫描功能,显示某行或列的灰度值,实现了对某些图像缺陷的动态检测。
With the cost of high-resolution, high-performance digital video camera coming down, an x-ray real-time imaging system was comprised by the group based on high-resolution digital camera and the new generation of image intensifier, which have the advantages of high spatial resolution and better anti-interference, compared with the system consisting of traditional image intensifier and ordinary analog video CCD camera. The paper was mainly to develop detection software, which contained image acquisition and processing functions, it can realize the calculation of characteristic parameters, but also be able to assist evaluation staff to determine the level of defects.
First of all, the hardware structure of the imaging system was introduced in the paper. After being familiar with the CCD digital camera and acquisition card, the image acquisition and preservation software was compiled, and then according to the request of x-ray detection and the actual needs of auto aluminum wheels detection, a suite of software was developed. In the image pre-processing part, by analyzing the image noise, the multi-frame superposition and median filtering method were respectively used to achieve the expected results. For the low contrast, gray-scale transformation, histogram equalization and other methods were adopted to enhance the image. Aiming at the characteristics of x-ray aluminum wheels image and the shortage of conventional image enhancement features, a new enhancement algorithm based on Curve let transform was presented, then the contrast and the quality of the original images were improved, and the better visual effects of the minutia area was achieved. In the image segmentation stage, the subtraction technique was introduced, and got a nice result, but in the segmentation process, it was likely to cause distortion defects. Then based on the analysis of the traditional edge detection algorithm, a combination algorithm of edge detection algorithm based on the SUSAN operator and target area was proposed, and achieved the desired result. Finally, on the basis of the testing requirement and casting defects level standard, defect classification, characteristics of expression and presentation methods were introduced, and then the perimeter, area and non-geometrical characteristics were measured. Furthermore, a real-time display system was designed, so that defect features were real-time displayed. In addition, the scanning function was also designed and a row or column of the gray value can be displayed, achieving the dynamic test to a certain image defect.
首先,介绍了课题组搭建的一套成像系统的硬件组成。在熟悉CCD数字相机及采集卡以后,编制了图像采集及保存软件,然后根据常规X射线检测的需要和汽车铝轮毂检测项目的实际需要,开发了一整套软件。在图像预处理部分,通过分析图像噪声,分别采用多帧叠加和中值滤波方法,达到了预期的结果。对于对比度不高的问题,采用灰度变换、直方图均衡等方法,对图像进行了增强；针对轮毂图像的特点及常规增强方法的不足,提出了一种基于Curve let变换的增强算法,提高了原图像的对比度,达到了细节区域较好的视觉效果。在图像分割阶段,采用减影技术进行分割,达到很好的分割效果,但是在分割过程中,容易造成缺陷的失真。然后在分析传统边缘提取算法的基础上,提出了基于SUSAN算子的边缘检测和基于目标区域相结合的算法,取得了理想的效果。结合企业对检测的要求及铸件缺陷评判标准,介绍了缺陷的分类、特征表达和表述方法,对周长、面积以及非几何特征进行测量,并设计了一种特征测量的实时显示系统,使缺陷特征能够实时显示出来。另外,还设计了扫描功能,显示某行或列的灰度值,实现了对某些图像缺陷的动态检测。
With the cost of high-resolution, high-performance digital video camera coming down, an x-ray real-time imaging system was comprised by the group based on high-resolution digital camera and the new generation of image intensifier, which have the advantages of high spatial resolution and better anti-interference, compared with the system consisting of traditional image intensifier and ordinary analog video CCD camera. The paper was mainly to develop detection software, which contained image acquisition and processing functions, it can realize the calculation of characteristic parameters, but also be able to assist evaluation staff to determine the level of defects.
First of all, the hardware structure of the imaging system was introduced in the paper. After being familiar with the CCD digital camera and acquisition card, the image acquisition and preservation software was compiled, and then according to the request of x-ray detection and the actual needs of auto aluminum wheels detection, a suite of software was developed. In the image pre-processing part, by analyzing the image noise, the multi-frame superposition and median filtering method were respectively used to achieve the expected results. For the low contrast, gray-scale transformation, histogram equalization and other methods were adopted to enhance the image. Aiming at the characteristics of x-ray aluminum wheels image and the shortage of conventional image enhancement features, a new enhancement algorithm based on Curve let transform was presented, then the contrast and the quality of the original images were improved, and the better visual effects of the minutia area was achieved. In the image segmentation stage, the subtraction technique was introduced, and got a nice result, but in the segmentation process, it was likely to cause distortion defects. Then based on the analysis of the traditional edge detection algorithm, a combination algorithm of edge detection algorithm based on the SUSAN operator and target area was proposed, and achieved the desired result. Finally, on the basis of the testing requirement and casting defects level standard, defect classification, characteristics of expression and presentation methods were introduced, and then the perimeter, area and non-geometrical characteristics were measured. Furthermore, a real-time display system was designed, so that defect features were real-time displayed. In addition, the scanning function was also designed and a row or column of the gray value can be displayed, achieving the dynamic test to a certain image defect.
引文
[1]刘福顺,汤明.无损检测基础[M].北京：北京航空航天大学出版社,2002：4～6.
[2]程耀瑜.工业射线实时成像检测技术研究及高性能数字成像系统研制[D].博士学位论文.江苏：南京理工大学,2003.
[3]曾祥照.无损检测文化概论[J].无损探伤,2002,26(2)：34-37.
[4]李国华,吴淼.现代无损检测与评价[M].北京：化学工业出版社,2009：15～17.
[5]李永红.高温熔体界面状态的X射线数字图像检测技术[D].硕士学位论文.山西：中北大学,2006.
[6]刘艳.X射线实时成像系统的发展[J].机械管理开发.2006,89(2)：7～8.
[7]何东健.数字图像处理[M].西安：西安电子科技大学出版社,2005：5～6.
[8]阮秋琦.数字图像处理学[M].北京：电子工业出版社,2004：9～10.
[9]陈树越,路宏年.数字式X射线成像无损检测技术[J].华北工学院学报,1999,20(1)：49～53.
[10]张晓光,高顶.射线检测焊接缺陷的提取和自动识别[M].北京：国防工业出版社,2004：22～25.
[11]李衍.射线实时成像检测最新欧洲标准[J].无损检测,2005,22(7)：50～53.
[12]李以善,刘德镇等.焊接结构检测技术[M].北京：化学工业出版社,2009：11～14.
[13]郑世才.射线照相检验技术的基础[J].无损检测,2002,22(1)：42～47.
[14]蔡文贵,李永远,许振华.CCD技术及应用[M].北京：电子工业出版社,1992：25～28.
[15]缪家鼎,徐文娟,牟同升.光电技术[M].杭州：浙江大学出版社,2006：17～19.
[16]M.Rossi,F.Casali,S.V.Golovkin,V.N.Govorun. Digital radiography using an EBCCD-based imaging device.Applied Radiation and Isotopes,2000,53:699～709.
[17]周长发.精通Visual C++图像处理编程[M].北京：电子工业出版社,2006：32～35.
[18]Richard C.Leinecker. Visual C++5.0开发技术内幕[M].陈冠民等译.北京：机械工业,1999：345～346.
[19]Kehoe A, Parker GA. Image Processing for Industrial Radiographic Inspection: Image Enhancement. British Journal of NDT,1990,32(3):183～190.
[20]Guylaine Daillant etc al.Defects in a weld:a complete Radiographic Processing Line. IEEE.1996:719～724.
[21]张兆礼,赵春晖,梅晓丹.现代图像处理技术及Matlab实现[M].北京：人民邮电出版社,2001：28～30.
[22]任大海,尤政.X射线检测中的图像质量的改善[J].仪表技术与传感器.2001(1)：36-37,41.
[23]Whiting B R,Muka E.Image quantization:statistics and modeling[J].Proc SPIE,1998,33(36):260-271.
[24]赵荣椿,赵忠明.数字图像处理导论[M].西安：西北工业大学出版社,2000：230～231.
[25]吴一全,王厚枢.图像对比度增强处理方法(一)[J].数据采集与处理,1989,(4)：39～49.
[26]夏良正.数字图像处理[M].南京：东南大学出版社,1999：103～104.
[27]Whiting B R.Signal statistics of X-ray computed tomography [J].Proc SPIE,2002 46(82):53～60.
[28]Just T,Thale W,Clausen R,et al.Interpretation of radiographs by digital image processing [J/OL].The E-Journal of Nondestructive Testing,1998,3(10):70～75.
[29]李庆丰,李巍,丁淑杰.数字图像分析技术在无损检测领域的应用及发展[J].黑龙江电子技术,1999(3)：58～59.
[30]何斌,马天予,王运坚等.Visual C++数字图像处理.第二版[M].北京：人民邮电出版社,2002：92～112.
[31]龚声容,刘纯平,王强等.数字图像处理与分析[M].北京：清华大学出版社,2006：6～9.
[32]William K.Pratt.数字图像处理[M].邓鲁华,张延恒译.北京：机械工业出版社,2005：198～223.
[33]贾永红.数字图像处理[M].武汉：武汉大学出版社,2003：86～98.
[34]程耀瑜,韩焱,潘德恒,等.高、低能X射线数字成像内视仪的研制[J].仪器仪表学报,2002,23(6)：579～583.
[35]E.J. Candes, David L Donoho. Curve let:a surprisingly effective nonadaptive representation of objects with edges. In Curve and Surface Fitting:Saint—Malo1999, Albert Cohen, Christophe Rabut, and Larry L. Schumaker, Eds. Vanderbilt University Press, Nashville. USA:department of Statistics, Stanford University,1999:965～976.
[36]Candes, E.J. and Donoho, D.L. Continuous Curve let Transform:Ⅱ Discretization into Frames. To appear Applied and Computational Harmonic Analysis http://www.acm.caltech.edu/～emmanuel/publications.html.2003:631～634.
[37]Candes,E.J. and Donoho, D.L. Continuous Curve let Transform:I Resolution of the Wave front Set. To appear Applied and Computational Harmonic Analysis http://www.acm.caltech.edu/～emmanuel/publications.html.2003:243～249.
[38]A. Grossman, J.Morlet. Decomposition of Hardy Functions into Square Integrable Wavelets of Constant Shape [J].SIAM J.Appl,1984,15:723-737.
[39]谭明金.Visual C++图形编程技巧与实例[M].北京：人民邮电出版社,2002：345～346.
[40]竺子民.光电图像处理[M].武汉：华中理工大学出版社,2000：57～58.
[41]贾云得.机器视觉[M].北京：科学出版社,2000：95～98.
[42]Pal S K, King R A. Image enhancement using fuzzy sets. Electronics Letters,1980, 16(10):376～378.
[43]C.Lehr,K.L.Feiste,D.Sregemann et al.Three Dimensional Defect Analysis Using Stereoradioscopy Based on Camera Modelling[C].Proceedings of the 7th ECNDT,1998:243-249.
[44]Domongo Mery,Dieter Filbert.Improvement in automated Aluminum Casting Inspection by Finding Correspondence of Potential Flaws in Multiple Radioscopic Images[C].Proceedingsof the 15th WCNDT,Rome,2000:631～645.
[45]艾民,阮兴云.数字减影的历史、现状和未来发展趋势[J].医疗装备,2000,13(6)：1～6.
[46]Puentes J,Roux C,Garreau,Metcal.Dynamic feature extraction of coronary artery motion using DSA image sequences. Medical Imaging, IEEE Transactions on,1998,17(6):857～871.
[47]Meijering E H W, Niessen W J, Viegever M A. Retrospective motion correction in digital subtraction angiography:areview. Medical Imaging, IEEE Transactions on,1999,18(1):2～21.
[48]周伟,余华民.图像处理与模式识别技术在焊缝射线检验中的应用[J].无损探伤.1990,14(2)：1～5.
[49]梁吉等.视觉检测系统及其应用[J].微计算机信息,2003,(1)：44～46.
[50]王瑞芳等.一种新型的视觉系统一原理和概念设计[J].光学技术,2001,(5)：420～423.
[51]中华人民共和国国家标准.金属熔化焊焊缝缺陷分类及说明.GB6417-86,1986.
[52]张毓晋.图像工程(上册)：图像处理和分析[M],北京：清华大学出版社,1999：260～262.
[53]孙即祥.图像分析[M].北京：科学出版社,2005：365～398.
[2]程耀瑜.工业射线实时成像检测技术研究及高性能数字成像系统研制[D].博士学位论文.江苏：南京理工大学,2003.
[3]曾祥照.无损检测文化概论[J].无损探伤,2002,26(2)：34-37.
[4]李国华,吴淼.现代无损检测与评价[M].北京：化学工业出版社,2009：15～17.
[5]李永红.高温熔体界面状态的X射线数字图像检测技术[D].硕士学位论文.山西：中北大学,2006.
[6]刘艳.X射线实时成像系统的发展[J].机械管理开发.2006,89(2)：7～8.
[7]何东健.数字图像处理[M].西安：西安电子科技大学出版社,2005：5～6.
[8]阮秋琦.数字图像处理学[M].北京：电子工业出版社,2004：9～10.
[9]陈树越,路宏年.数字式X射线成像无损检测技术[J].华北工学院学报,1999,20(1)：49～53.
[10]张晓光,高顶.射线检测焊接缺陷的提取和自动识别[M].北京：国防工业出版社,2004：22～25.
[11]李衍.射线实时成像检测最新欧洲标准[J].无损检测,2005,22(7)：50～53.
[12]李以善,刘德镇等.焊接结构检测技术[M].北京：化学工业出版社,2009：11～14.
[13]郑世才.射线照相检验技术的基础[J].无损检测,2002,22(1)：42～47.
[14]蔡文贵,李永远,许振华.CCD技术及应用[M].北京：电子工业出版社,1992：25～28.
[15]缪家鼎,徐文娟,牟同升.光电技术[M].杭州：浙江大学出版社,2006：17～19.
[16]M.Rossi,F.Casali,S.V.Golovkin,V.N.Govorun. Digital radiography using an EBCCD-based imaging device.Applied Radiation and Isotopes,2000,53:699～709.
[17]周长发.精通Visual C++图像处理编程[M].北京：电子工业出版社,2006：32～35.
[18]Richard C.Leinecker. Visual C++5.0开发技术内幕[M].陈冠民等译.北京：机械工业,1999：345～346.
[19]Kehoe A, Parker GA. Image Processing for Industrial Radiographic Inspection: Image Enhancement. British Journal of NDT,1990,32(3):183～190.
[20]Guylaine Daillant etc al.Defects in a weld:a complete Radiographic Processing Line. IEEE.1996:719～724.
[21]张兆礼,赵春晖,梅晓丹.现代图像处理技术及Matlab实现[M].北京：人民邮电出版社,2001：28～30.
[22]任大海,尤政.X射线检测中的图像质量的改善[J].仪表技术与传感器.2001(1)：36-37,41.
[23]Whiting B R,Muka E.Image quantization:statistics and modeling[J].Proc SPIE,1998,33(36):260-271.
[24]赵荣椿,赵忠明.数字图像处理导论[M].西安：西北工业大学出版社,2000：230～231.
[25]吴一全,王厚枢.图像对比度增强处理方法(一)[J].数据采集与处理,1989,(4)：39～49.
[26]夏良正.数字图像处理[M].南京：东南大学出版社,1999：103～104.
[27]Whiting B R.Signal statistics of X-ray computed tomography [J].Proc SPIE,2002 46(82):53～60.
[28]Just T,Thale W,Clausen R,et al.Interpretation of radiographs by digital image processing [J/OL].The E-Journal of Nondestructive Testing,1998,3(10):70～75.
[29]李庆丰,李巍,丁淑杰.数字图像分析技术在无损检测领域的应用及发展[J].黑龙江电子技术,1999(3)：58～59.
[30]何斌,马天予,王运坚等.Visual C++数字图像处理.第二版[M].北京：人民邮电出版社,2002：92～112.
[31]龚声容,刘纯平,王强等.数字图像处理与分析[M].北京：清华大学出版社,2006：6～9.
[32]William K.Pratt.数字图像处理[M].邓鲁华,张延恒译.北京：机械工业出版社,2005：198～223.
[33]贾永红.数字图像处理[M].武汉：武汉大学出版社,2003：86～98.
[34]程耀瑜,韩焱,潘德恒,等.高、低能X射线数字成像内视仪的研制[J].仪器仪表学报,2002,23(6)：579～583.
[35]E.J. Candes, David L Donoho. Curve let:a surprisingly effective nonadaptive representation of objects with edges. In Curve and Surface Fitting:Saint—Malo1999, Albert Cohen, Christophe Rabut, and Larry L. Schumaker, Eds. Vanderbilt University Press, Nashville. USA:department of Statistics, Stanford University,1999:965～976.
[36]Candes, E.J. and Donoho, D.L. Continuous Curve let Transform:Ⅱ Discretization into Frames. To appear Applied and Computational Harmonic Analysis http://www.acm.caltech.edu/～emmanuel/publications.html.2003:631～634.
[37]Candes,E.J. and Donoho, D.L. Continuous Curve let Transform:I Resolution of the Wave front Set. To appear Applied and Computational Harmonic Analysis http://www.acm.caltech.edu/～emmanuel/publications.html.2003:243～249.
[38]A. Grossman, J.Morlet. Decomposition of Hardy Functions into Square Integrable Wavelets of Constant Shape [J].SIAM J.Appl,1984,15:723-737.
[39]谭明金.Visual C++图形编程技巧与实例[M].北京：人民邮电出版社,2002：345～346.
[40]竺子民.光电图像处理[M].武汉：华中理工大学出版社,2000：57～58.
[41]贾云得.机器视觉[M].北京：科学出版社,2000：95～98.
[42]Pal S K, King R A. Image enhancement using fuzzy sets. Electronics Letters,1980, 16(10):376～378.
[43]C.Lehr,K.L.Feiste,D.Sregemann et al.Three Dimensional Defect Analysis Using Stereoradioscopy Based on Camera Modelling[C].Proceedings of the 7th ECNDT,1998:243-249.
[44]Domongo Mery,Dieter Filbert.Improvement in automated Aluminum Casting Inspection by Finding Correspondence of Potential Flaws in Multiple Radioscopic Images[C].Proceedingsof the 15th WCNDT,Rome,2000:631～645.
[45]艾民,阮兴云.数字减影的历史、现状和未来发展趋势[J].医疗装备,2000,13(6)：1～6.
[46]Puentes J,Roux C,Garreau,Metcal.Dynamic feature extraction of coronary artery motion using DSA image sequences. Medical Imaging, IEEE Transactions on,1998,17(6):857～871.
[47]Meijering E H W, Niessen W J, Viegever M A. Retrospective motion correction in digital subtraction angiography:areview. Medical Imaging, IEEE Transactions on,1999,18(1):2～21.
[48]周伟,余华民.图像处理与模式识别技术在焊缝射线检验中的应用[J].无损探伤.1990,14(2)：1～5.
[49]梁吉等.视觉检测系统及其应用[J].微计算机信息,2003,(1)：44～46.
[50]王瑞芳等.一种新型的视觉系统一原理和概念设计[J].光学技术,2001,(5)：420～423.
[51]中华人民共和国国家标准.金属熔化焊焊缝缺陷分类及说明.GB6417-86,1986.
[52]张毓晋.图像工程(上册)：图像处理和分析[M],北京：清华大学出版社,1999：260～262.
[53]孙即祥.图像分析[M].北京：科学出版社,2005：365～398.