轮胎X光图像自动识别系统算法研究
|
摘要
使用X光机对轮胎进行检测，已作为质量控制的有效手段被国内外的厂商广泛采用和证明。在国内长期以来，人们一直是凭肉眼对轮胎图像进行检测，此方法重复一致性差、人力资源耗费量大、无法很好的满足企业信息化管理的要求。本文所作的工作就是要研究出一套算法，可根据轮胎的X光图像对轮胎的各种缺陷进行自动判别，实现无人值守的检测，解决人眼检测所带来的一系列问题。目前，国外的厂商已推出了产品化的全自动轮胎检测系统，而国内却没有此类的研究，我们的工作将填补国内该领域的研究空白，并致力于实现轮胎自动检测产品的国产化。
本文主要研究了实现自动检测各种轮胎缺陷的一系列算法。首先对得到的轮胎图像进行一系列预处理，包括检测窗口的选取、图像增强等。然后，本文提出一种改进的基于一维投影的快速匹配算法，将其应用到轮胎缺陷的自动识别中去，并提出一种新的基于局部不变特征的直方图相关匹配算法，通过此方法完成对边部劈缝的自动识别。接下来，针对可能存在缺陷的胎侧部位图像，利用提取到的图像特征及各种缺陷的自身特点来对缺陷进行定性、定量分析。最后对轮胎质量给出判断结果，并把存在缺陷部位的图像在图中加以标记显示。
研究中使用Visual C#.NET语言搭建了测试平台，以大量的实际图片进行了仿真试验，得到了较为令人满意的结果。文中给出了针对各个轮胎缺陷的仿真结果及相关数据。
经实验证明，本文提出的轮胎各种缺陷的自动检测算法是有效的，它可寻找出所有存在缺陷的部位，并可对部分常见缺陷的种类及缺陷程度给出较为准确的判断。
It is broadly proved that X-ray is efficient in tire inspection in order control the quality. Inspect tires using naked eyes is long-standing at home, which could make poor consistent and repeatable result, manpower waste and could make it hard to information manage. We propose a novel algorithm to solve such questions and make the inspection procedure without operator intervention. At present overseas manufactures have already developed tire automatic inspection system, but there is none such research at home. Our job would full up the blank in that field.
This paper mostly studied how to inspect tires' disfigurement automatically by using the tire X-ray image. Firstly, preprocessing the image, such as select the interesting area and image enhancement. And then, we proposed a novel one dimension projective based fast image matching algorithm to inspect tire disfigurements. A novel image matching algorithm based on local invariant feature and histogram-based similar distance is also proposed to inspect dog ears. The next, aim at the area which probably contains disfigurements, take use of the extracted features and the features each disfigurement has to determine which kind of disfigurement it is. Finally, estimate the degree of the tires, and mark the area which the disfigurement exists.
A great amount of images are used to test the algorithm based on the framework programmed by Visual C#.NET, and all get satisfactory results. The emulation results of each disfigurement are presented in this paper.
It is proved that the algorithm we proposed is efficient to inspect the tire without operator intervention. Some of the common disfigurements are well recognized.
本文主要研究了实现自动检测各种轮胎缺陷的一系列算法。首先对得到的轮胎图像进行一系列预处理，包括检测窗口的选取、图像增强等。然后，本文提出一种改进的基于一维投影的快速匹配算法，将其应用到轮胎缺陷的自动识别中去，并提出一种新的基于局部不变特征的直方图相关匹配算法，通过此方法完成对边部劈缝的自动识别。接下来，针对可能存在缺陷的胎侧部位图像，利用提取到的图像特征及各种缺陷的自身特点来对缺陷进行定性、定量分析。最后对轮胎质量给出判断结果，并把存在缺陷部位的图像在图中加以标记显示。
研究中使用Visual C#.NET语言搭建了测试平台，以大量的实际图片进行了仿真试验，得到了较为令人满意的结果。文中给出了针对各个轮胎缺陷的仿真结果及相关数据。
经实验证明，本文提出的轮胎各种缺陷的自动检测算法是有效的，它可寻找出所有存在缺陷的部位，并可对部分常见缺陷的种类及缺陷程度给出较为准确的判断。
It is broadly proved that X-ray is efficient in tire inspection in order control the quality. Inspect tires using naked eyes is long-standing at home, which could make poor consistent and repeatable result, manpower waste and could make it hard to information manage. We propose a novel algorithm to solve such questions and make the inspection procedure without operator intervention. At present overseas manufactures have already developed tire automatic inspection system, but there is none such research at home. Our job would full up the blank in that field.
This paper mostly studied how to inspect tires' disfigurement automatically by using the tire X-ray image. Firstly, preprocessing the image, such as select the interesting area and image enhancement. And then, we proposed a novel one dimension projective based fast image matching algorithm to inspect tire disfigurements. A novel image matching algorithm based on local invariant feature and histogram-based similar distance is also proposed to inspect dog ears. The next, aim at the area which probably contains disfigurements, take use of the extracted features and the features each disfigurement has to determine which kind of disfigurement it is. Finally, estimate the degree of the tires, and mark the area which the disfigurement exists.
A great amount of images are used to test the algorithm based on the framework programmed by Visual C#.NET, and all get satisfactory results. The emulation results of each disfigurement are presented in this paper.
It is proved that the algorithm we proposed is efficient to inspect the tire without operator intervention. Some of the common disfigurements are well recognized.
引文
[1] 孙远，周刚慧，赵立初等．灰度图像匹配的快速算法[J]．上海交通大学学报，2000，5：702-704
[2] Lei Qin, Wen Gao. Image Matching Based on A Local Invariant Descriptor[C]. IEEE International conference on Image Processing,2005, 3: 377-380
[3] Rafael C. Gonzalez, Richard E. Woods. 数字图像处理(第二版)[M]，北京：电子工业出版社，2005：540-543
[4] 贾渊，姬长英．采用改进的自适应模糊多级中值滤波算法去除牛肉图像斑点噪声[J]．农业工程学报，2005，21(2)：102-105
[5] Marcos C. De Andrade, Baba C. Vemuri. An interactive algorithm for image denoising and segmentation[J]. IEEE Trans on Image Processing, 2001 : 274-281
[6] Hao Zhao, Weidong Zhou, Yuhua Peng. Application of independent component anlysis on noisy image separation[C]. IEEE ICSP'04 proceedings, 2004:1018-1021
[7] 费佩燕，郭宝龙．二进小波变换的图像去噪[J]．计算机工程与运用，2003，8：9-12
[8] 李春鑫，李天伟，王孝通，黄谦等．改进小波包分析在图像噪声处理中的应用[J]．计算机工程与设计，2005，26(9)：2550-2554
[9] D. L. Donoho. De-noising via soft-thresholding [J].IEEE Trans on Info Theo,1992,41 (3):613-627
[10] D.L.Donoho, I.M.Johnstone. Ideal spatial adaptation via wavelet shinkage[J]. Biometrika, 1994,81 (2):425-455
[11] D.L.Donoho, I.M.Johnstone. Adapting to unknown smoothness via wavelet shrinkage[J]. J. Amer. Stat. Assoc.,1995,90:1200-1224
[12] I. M. Johnstone, B. W. Siiverman. Wavelet threshold estimators for data with correlated noise[J].J.Roy.Stat.Soc.B,1997,59:319-351
[13] T.Cai. Adaptive wavelet estimation :a block thresholding and oracle inequality approach[J].Ann.Stat., 1999,27:898-924
[14] P.Hall, G.Kerkyacharian ,and D.Picard. On the minimax optimality of block thresholded wavelet estimators[J].Stat.Sinica, 1999,9:33-50
[15] G. Chen, T. Bui. Mutiwavelets denoising using neighboring coefficients[J].IEEE Signal Process. Lett.,2003,10 (7):211-214
[16] Mallat S. Multifrequency channel decompositions[J].IEEE Trans on ASSP,1989,37 (12):2091-2110
[17] Kenneth. R. Castleman．数字图像处理[M]。北京：电子工业出版社，1998：359-368
[18] D.L.Donoho, I.M.Johnstone, G. Kerkyacharian,D.Picard. Wavelet shrinkage: asymptopia?,??Journal of the Royal Statistical Society B,1995,57:301-369
[19] 沈庭芝，方子文．数字图像处理及模式识别[M]．北京：北京理工大学出版社，1998：150-151
[20] Huttenlocher D P, Rucklidge W J. Comparing images using the Hausdorff distance under translating[C].IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1992
[21] Fookes C, Bennamoun M. The use of mutual information for rigid medical image registration : A review[C]. IEEE International Conference on Systems, Man and Cybemetics,2002
[22] 廖云涛，任仙怡，张桂林，等．一种新的基于对应像素距离度量的图像相关匹配方法[J]．红外与激光工程，2001，30(6)：418-421
[23] 王红梅，张科，李言俊．图像匹配研究进展[J]．计算机工程与应用，2004，19：42-44，77
[24] Morave H.P.Towards automatic visual obstacle avoidance [C].In:Proceeding of the 5th International Joint Conference on Artificial Intelligence, 1977
[25] Harris C, Stephens M A. Combined corner and edge detector[C].ln:Proceedings of the 4th Alvey Vision Conference, 1988
[26] Vincent T, Laganiere R. Matching feature points for telerobotics[C].In: 1~(st) International Workshop on HAVE and their Applications,2002
[27] You J, Bhattacharya P A. Wavelet-based coarse-to-fine image matching scheme in a parallel virtual machine environment[J].IEEE Trans on Image Processing, 2000, 9(9): 1547-1559
[28] Hu M K. Pattern recognition by moments invariants[C].In:Proc IRE,1961
[29] Zeng zhanggui, Yan Hong. Region matching and optimal matching pair theorem[C].ln: Computer Graphics International 2001
[30] Huang zhaohui, Cohen F S. Affine-invariant B-Spline moments for curve matching[J].IEEE Trans on Image Processing,1996, 5 (10): 1473-1480
[31] 刘雅轩，苏秀琴，王萍．一种基于局部投影熵的图像匹配新算法[J]．光子学报，2004，33(1)：105-108
[32] 郭显久，李莉．基于细分小波与局部投影熵相结合的图像匹配算法[J]．辽宁师范大学学报(自然科学版)，2005，28(1)：59-63
[33] 胡敏，丁益洪，平西建．基于属性关系直方图统计的线状纹理图像检索方法[J]．计算机辅助设计与图形学学报，15(1)：48-52
[34] C. Schmid, R. Mohr. Local Grayvalue lnvariants for Image Retrieval[J]. IEEE PAMI, 1997, 19: 530-534
[35] D. G. Lowe. Distinctive Image Features from Scale-Invariant Keypoints[J]. International Journal ofComputer Vision, 2004, 60(2): 91-110[36] C. Harris, M. Stephens. A Combined Comer and Edge Detector[C]. In: Proc. Alvey Vision Conf, 1988:189-192
[37] T. Tuytelaars, L. Van Gool. Wide Baseline Stereo Matching Based on Local Affinely Invariant Regions[C]. In: BMVC, 2000:412-425
[38] 张桂林，徐捷．频域相关技术在图像匹配中的应用[J]．模式识别与人工智能，1997，10(1)：87-92
[39] 李静．字符串的模式匹配算法——基于KMP算法的讨论[J]．青岛化工学院学报，2002，6：78-80
[40] Andrew S. Glassner. Graphics Gems [M]. Academic Press Inc, (1995) , 327-329
[41] D. P. Huttonlocher, G. A. Klandeman, W. J. Ruchlidge, Comparing images using the Hausdorff distance[J].IEEE Trans. on Pattern Analysis and Machine Intelligence, 1993, 15 (9): 850-863
[42] 王慧燕．图像边缘检测和图像匹配研究及应用(博士论文)．浙江大学，2003
[43] 黄勇，王建国，黄顺吉．SAR图像匹配算法及实现．信号处理，2003，19(2)：79-181，156
[44] 韩建栋，向健勇，刘畅．一种新的边缘检测算法[J]．电子科技，2004，3：53-55
[45] 张燕，曾立波，吴琼水，谢文娟．一种适用于任意形状区域的快速空洞填充算法[J]．计算机应用研究，2004，12：155-156，136
[46] 钟志光，卢君，刘伟荣．Visual C++．NET数字图像处理实例与解析[M]．北京：清华大学出版社，2003：512-529
[47] 俞妍妍，王继成．改进的多尺度形态形态梯度的图像边缘检测[J]．计算机工程与应用，2003，18：75-82
[48] 杨照华，浦昭邦，祁振强．保持图像细节的多尺度形态滤波的新方法[J]．光电子·激光，2003，14(8)：862-865
[49] 周焰，王祖喜，李德华．基于弧长参数的曲线自适应平滑算法[J]．红外与激光工程，2000，1：25-29
[50] 尹平，王润生．自适应多尺度边缘检测[J]．软件学报，2000，11(8)：990-994
[51] 钟志光，卢君，刘伟荣．Visual C++．NET数字图像处理实例与解析[M]．北京：清华大学出版社，2003：504-512
[52] 井上诚喜，八木伸行，林正树等．C语言实用数字图像处理[M]．北京：科学出版社，2003：81-88
[53] Alistair J. Fitch, Alexander Kadyrov, William J. Christmas etc. Fast Robust Correlation. IEEE Trans on Image Processing,2005,14 (8) :1063-1073
[54] 罗玲，解梅，陈杉．基于多尺度形态滤波的分水岭图像分割方法[J]．计算机辅助设计与图形学学报，2004，2：168-173
[55] 罗文村．基于阈值法与区域生长法综合集成的图像分割法[J]．现代计算机，2001，5：??43-46
[56] 魏弘博，吕振肃，蒋田仔，刘新艳．图像分割技术纵览[J]．甘肃科学学报，2004，6：19-24
[57] 徐慧．数字图像实用工程案例精选[M]．北京：人民邮电出版社，2004
[58] 曹炬，马杰，谭毅华，田金文．基于像素抽样的快速互相关图像匹配算法[J]．宇航学报，2004．3：173-178
[59] S. Saitoh. Theory of reproducing kernels and its applications[M]. Longman Scientific & Technical Press,1988:1-15
[60] 钟志光，卢君，刘伟荣．Visual C++．NET数字图像处理实例与解析[M]．北京：清华大学出版社，2003：403-405
[2] Lei Qin, Wen Gao. Image Matching Based on A Local Invariant Descriptor[C]. IEEE International conference on Image Processing,2005, 3: 377-380
[3] Rafael C. Gonzalez, Richard E. Woods. 数字图像处理(第二版)[M]，北京：电子工业出版社，2005：540-543
[4] 贾渊，姬长英．采用改进的自适应模糊多级中值滤波算法去除牛肉图像斑点噪声[J]．农业工程学报，2005，21(2)：102-105
[5] Marcos C. De Andrade, Baba C. Vemuri. An interactive algorithm for image denoising and segmentation[J]. IEEE Trans on Image Processing, 2001 : 274-281
[6] Hao Zhao, Weidong Zhou, Yuhua Peng. Application of independent component anlysis on noisy image separation[C]. IEEE ICSP'04 proceedings, 2004:1018-1021
[7] 费佩燕，郭宝龙．二进小波变换的图像去噪[J]．计算机工程与运用，2003，8：9-12
[8] 李春鑫，李天伟，王孝通，黄谦等．改进小波包分析在图像噪声处理中的应用[J]．计算机工程与设计，2005，26(9)：2550-2554
[9] D. L. Donoho. De-noising via soft-thresholding [J].IEEE Trans on Info Theo,1992,41 (3):613-627
[10] D.L.Donoho, I.M.Johnstone. Ideal spatial adaptation via wavelet shinkage[J]. Biometrika, 1994,81 (2):425-455
[11] D.L.Donoho, I.M.Johnstone. Adapting to unknown smoothness via wavelet shrinkage[J]. J. Amer. Stat. Assoc.,1995,90:1200-1224
[12] I. M. Johnstone, B. W. Siiverman. Wavelet threshold estimators for data with correlated noise[J].J.Roy.Stat.Soc.B,1997,59:319-351
[13] T.Cai. Adaptive wavelet estimation :a block thresholding and oracle inequality approach[J].Ann.Stat., 1999,27:898-924
[14] P.Hall, G.Kerkyacharian ,and D.Picard. On the minimax optimality of block thresholded wavelet estimators[J].Stat.Sinica, 1999,9:33-50
[15] G. Chen, T. Bui. Mutiwavelets denoising using neighboring coefficients[J].IEEE Signal Process. Lett.,2003,10 (7):211-214
[16] Mallat S. Multifrequency channel decompositions[J].IEEE Trans on ASSP,1989,37 (12):2091-2110
[17] Kenneth. R. Castleman．数字图像处理[M]。北京：电子工业出版社，1998：359-368
[18] D.L.Donoho, I.M.Johnstone, G. Kerkyacharian,D.Picard. Wavelet shrinkage: asymptopia?,??Journal of the Royal Statistical Society B,1995,57:301-369
[19] 沈庭芝，方子文．数字图像处理及模式识别[M]．北京：北京理工大学出版社，1998：150-151
[20] Huttenlocher D P, Rucklidge W J. Comparing images using the Hausdorff distance under translating[C].IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1992
[21] Fookes C, Bennamoun M. The use of mutual information for rigid medical image registration : A review[C]. IEEE International Conference on Systems, Man and Cybemetics,2002
[22] 廖云涛，任仙怡，张桂林，等．一种新的基于对应像素距离度量的图像相关匹配方法[J]．红外与激光工程，2001，30(6)：418-421
[23] 王红梅，张科，李言俊．图像匹配研究进展[J]．计算机工程与应用，2004，19：42-44，77
[24] Morave H.P.Towards automatic visual obstacle avoidance [C].In:Proceeding of the 5th International Joint Conference on Artificial Intelligence, 1977
[25] Harris C, Stephens M A. Combined corner and edge detector[C].ln:Proceedings of the 4th Alvey Vision Conference, 1988
[26] Vincent T, Laganiere R. Matching feature points for telerobotics[C].In: 1~(st) International Workshop on HAVE and their Applications,2002
[27] You J, Bhattacharya P A. Wavelet-based coarse-to-fine image matching scheme in a parallel virtual machine environment[J].IEEE Trans on Image Processing, 2000, 9(9): 1547-1559
[28] Hu M K. Pattern recognition by moments invariants[C].In:Proc IRE,1961
[29] Zeng zhanggui, Yan Hong. Region matching and optimal matching pair theorem[C].ln: Computer Graphics International 2001
[30] Huang zhaohui, Cohen F S. Affine-invariant B-Spline moments for curve matching[J].IEEE Trans on Image Processing,1996, 5 (10): 1473-1480
[31] 刘雅轩，苏秀琴，王萍．一种基于局部投影熵的图像匹配新算法[J]．光子学报，2004，33(1)：105-108
[32] 郭显久，李莉．基于细分小波与局部投影熵相结合的图像匹配算法[J]．辽宁师范大学学报(自然科学版)，2005，28(1)：59-63
[33] 胡敏，丁益洪，平西建．基于属性关系直方图统计的线状纹理图像检索方法[J]．计算机辅助设计与图形学学报，15(1)：48-52
[34] C. Schmid, R. Mohr. Local Grayvalue lnvariants for Image Retrieval[J]. IEEE PAMI, 1997, 19: 530-534
[35] D. G. Lowe. Distinctive Image Features from Scale-Invariant Keypoints[J]. International Journal ofComputer Vision, 2004, 60(2): 91-110[36] C. Harris, M. Stephens. A Combined Comer and Edge Detector[C]. In: Proc. Alvey Vision Conf, 1988:189-192
[37] T. Tuytelaars, L. Van Gool. Wide Baseline Stereo Matching Based on Local Affinely Invariant Regions[C]. In: BMVC, 2000:412-425
[38] 张桂林，徐捷．频域相关技术在图像匹配中的应用[J]．模式识别与人工智能，1997，10(1)：87-92
[39] 李静．字符串的模式匹配算法——基于KMP算法的讨论[J]．青岛化工学院学报，2002，6：78-80
[40] Andrew S. Glassner. Graphics Gems [M]. Academic Press Inc, (1995) , 327-329
[41] D. P. Huttonlocher, G. A. Klandeman, W. J. Ruchlidge, Comparing images using the Hausdorff distance[J].IEEE Trans. on Pattern Analysis and Machine Intelligence, 1993, 15 (9): 850-863
[42] 王慧燕．图像边缘检测和图像匹配研究及应用(博士论文)．浙江大学，2003
[43] 黄勇，王建国，黄顺吉．SAR图像匹配算法及实现．信号处理，2003，19(2)：79-181，156
[44] 韩建栋，向健勇，刘畅．一种新的边缘检测算法[J]．电子科技，2004，3：53-55
[45] 张燕，曾立波，吴琼水，谢文娟．一种适用于任意形状区域的快速空洞填充算法[J]．计算机应用研究，2004，12：155-156，136
[46] 钟志光，卢君，刘伟荣．Visual C++．NET数字图像处理实例与解析[M]．北京：清华大学出版社，2003：512-529
[47] 俞妍妍，王继成．改进的多尺度形态形态梯度的图像边缘检测[J]．计算机工程与应用，2003，18：75-82
[48] 杨照华，浦昭邦，祁振强．保持图像细节的多尺度形态滤波的新方法[J]．光电子·激光，2003，14(8)：862-865
[49] 周焰，王祖喜，李德华．基于弧长参数的曲线自适应平滑算法[J]．红外与激光工程，2000，1：25-29
[50] 尹平，王润生．自适应多尺度边缘检测[J]．软件学报，2000，11(8)：990-994
[51] 钟志光，卢君，刘伟荣．Visual C++．NET数字图像处理实例与解析[M]．北京：清华大学出版社，2003：504-512
[52] 井上诚喜，八木伸行，林正树等．C语言实用数字图像处理[M]．北京：科学出版社，2003：81-88
[53] Alistair J. Fitch, Alexander Kadyrov, William J. Christmas etc. Fast Robust Correlation. IEEE Trans on Image Processing,2005,14 (8) :1063-1073
[54] 罗玲，解梅，陈杉．基于多尺度形态滤波的分水岭图像分割方法[J]．计算机辅助设计与图形学学报，2004，2：168-173
[55] 罗文村．基于阈值法与区域生长法综合集成的图像分割法[J]．现代计算机，2001，5：??43-46
[56] 魏弘博，吕振肃，蒋田仔，刘新艳．图像分割技术纵览[J]．甘肃科学学报，2004，6：19-24
[57] 徐慧．数字图像实用工程案例精选[M]．北京：人民邮电出版社，2004
[58] 曹炬，马杰，谭毅华，田金文．基于像素抽样的快速互相关图像匹配算法[J]．宇航学报，2004．3：173-178
[59] S. Saitoh. Theory of reproducing kernels and its applications[M]. Longman Scientific & Technical Press,1988:1-15
[60] 钟志光，卢君，刘伟荣．Visual C++．NET数字图像处理实例与解析[M]．北京：清华大学出版社，2003：403-405