旋转不变的圆形图像匹配方法研究
|
摘要
图像匹配技术是图像处理技术与机器视觉检测中的核心内容,在工业产品在线检测系统中,传统的匹配方法不适用于发生角度旋转的两幅图像间的匹配,而圆形图像在线检测系统中,采集到的图像与基准图像之间总是存在任意角度偏差,这给检测工作带来了一定的困难性。本课题正是基于这种背景下提出的,以啤酒瓶盖质量检测为例,针对存在角度旋转的圆形图像匹配进行了研究。
本文在分析旋转的圆形图像匹配方法发展现状及发展趋势的基础上,针对圆形图像的预处理以及匹配等工作进行了深入研究,主要研究内容如下：
在预处理阶段,本文对采集到的瓶盖图像的定位进行了研究,对图像进行分段线性变换并根据三点定位圆的数学理论完成圆形图像的定位。
针对图像旋转角度的计算问题,课题分别提出基于关键点检测的旋转角度校正算法与基于傅里叶变换的旋转角度校正算法。前者首先采用基于曲率尺度空间的角点检测算法检测角点,在已检测的角点中找到一个较稳定的关键点,以此关键点为特征点,将该点与图像的几何中心点连线,求出该线与坐标轴的夹角即为待测图像的旋转角度。后者利用傅里叶变换算法在时频域中的旋转不变性,以图像中心为原点,1/2图像宽度为半径做圆,统计在0-180度内径向方向上的灰度累加值,则实际旋转的角度即为最大值所对应的角度。两种算法可以实现相同的功能,但后者对旋转角度的计算具有更高的精度。
最后课题在对待测图像旋转校正的基础上,采用基于分块均值的快速块匹配算法和基于对数极坐标变换的匹配算法完成待测图像与基准图像之间的匹配。
本文根据以上所提出的方法进行了大量的实验,从解决旋转变化、匹配速度和正确率方面进行分析。实验结果表明,所提出的算法具有较高的匹配速度和鲁棒性,并且算法对于解决图像旋转角度的校正问题具有较好的优势和有效性。
The matching technology on image is the core content in picture processing technology and machine vision detecting. In the system of the online measuring of industrial products, the traditional matching method is not applicable to angle rotation between the two matching images, While in round images on-line detection system, there is always existence some angle deviation between the collecting images and benchmark, which brings some difficulties to the detection work. The issue is proposed based on this background, with the example of beer cap quality inspection, aiming to rotation changes of angle between two images this paper was studied.
This articles has done some research about the collected round image, and has carried out depth study for the preprocessing and matching of round image. The main research contents as follows.
In the pretreatment stage, the paper has made some studies about the location for the collected cap image. Making the round image isolated from the background which based on the mathematics theory of three points positing a circular by segmented linear transformation.
Then proposed two angle correction algorithms for image rotation that based on key point detection and Fourier transform. Using the adaptive corner detector based on curvature scale space algorithm to detect corner points, to select a stable key point in the corners which have been detected. Connection the selected point with the geometric center of the image in order to certain the angle of the rotation image. The latter Fourier transform algorithm was introduced, training its time-frequency domain as invariance to rotation, firstly the image rotation angle was corrected by using Fourier transform, which made the untested image and reference image in the same direction.
At last using the fast matching algorithm that based on sub block mean and the method of round image matching based on log-polar transform to match the untested and reference image.
The experimental results show that compared with existing matching algorithms, the proposed algorithm can solve the rotation with any angle between two images' matching, and the algorithm has a high matching accuracy and applicability. In addition, the validity and reliability had got verified.
本文在分析旋转的圆形图像匹配方法发展现状及发展趋势的基础上,针对圆形图像的预处理以及匹配等工作进行了深入研究,主要研究内容如下：
在预处理阶段,本文对采集到的瓶盖图像的定位进行了研究,对图像进行分段线性变换并根据三点定位圆的数学理论完成圆形图像的定位。
针对图像旋转角度的计算问题,课题分别提出基于关键点检测的旋转角度校正算法与基于傅里叶变换的旋转角度校正算法。前者首先采用基于曲率尺度空间的角点检测算法检测角点,在已检测的角点中找到一个较稳定的关键点,以此关键点为特征点,将该点与图像的几何中心点连线,求出该线与坐标轴的夹角即为待测图像的旋转角度。后者利用傅里叶变换算法在时频域中的旋转不变性,以图像中心为原点,1/2图像宽度为半径做圆,统计在0-180度内径向方向上的灰度累加值,则实际旋转的角度即为最大值所对应的角度。两种算法可以实现相同的功能,但后者对旋转角度的计算具有更高的精度。
最后课题在对待测图像旋转校正的基础上,采用基于分块均值的快速块匹配算法和基于对数极坐标变换的匹配算法完成待测图像与基准图像之间的匹配。
本文根据以上所提出的方法进行了大量的实验,从解决旋转变化、匹配速度和正确率方面进行分析。实验结果表明,所提出的算法具有较高的匹配速度和鲁棒性,并且算法对于解决图像旋转角度的校正问题具有较好的优势和有效性。
The matching technology on image is the core content in picture processing technology and machine vision detecting. In the system of the online measuring of industrial products, the traditional matching method is not applicable to angle rotation between the two matching images, While in round images on-line detection system, there is always existence some angle deviation between the collecting images and benchmark, which brings some difficulties to the detection work. The issue is proposed based on this background, with the example of beer cap quality inspection, aiming to rotation changes of angle between two images this paper was studied.
This articles has done some research about the collected round image, and has carried out depth study for the preprocessing and matching of round image. The main research contents as follows.
In the pretreatment stage, the paper has made some studies about the location for the collected cap image. Making the round image isolated from the background which based on the mathematics theory of three points positing a circular by segmented linear transformation.
Then proposed two angle correction algorithms for image rotation that based on key point detection and Fourier transform. Using the adaptive corner detector based on curvature scale space algorithm to detect corner points, to select a stable key point in the corners which have been detected. Connection the selected point with the geometric center of the image in order to certain the angle of the rotation image. The latter Fourier transform algorithm was introduced, training its time-frequency domain as invariance to rotation, firstly the image rotation angle was corrected by using Fourier transform, which made the untested image and reference image in the same direction.
At last using the fast matching algorithm that based on sub block mean and the method of round image matching based on log-polar transform to match the untested and reference image.
The experimental results show that compared with existing matching algorithms, the proposed algorithm can solve the rotation with any angle between two images' matching, and the algorithm has a high matching accuracy and applicability. In addition, the validity and reliability had got verified.
引文
[1]刘止.图像实时匹配算法及其应用研究：(硕士学位论文).哈尔滨：哈尔滨工业大学,2009.
[2]Ehlers.M, Fuller M. Region-based matching for image integration in remote sensing databases. Dig Int Geosci mote Sens Symp (IGARSS):1991(4):2231-2234.
[3]Brown L G.. Survey of image registration techniques. ACM Computing Surveys:1992,24(4): 325-376.
[4]Sim D G, Kim H K, etal. Translation scale, and rotation in variant texture descriptor for texture-based image retrieval. IEEE Int ConfImage Process:2000,3:742-745.
[5]Farhan Ullah, Shunichi Kaneko. Using orientation codes for rota-tion-invariant template matching. Pattern Recognition:2004.
[6]Sang-Woo Lee, Whoi-Yul. Kim. Rotation-invariant template matching using projection method. Proc KITE:1996,19(1):475-476.
[7]王钰,孙小惟.基于圆投影及直方图不变特征的图像匹配方法.模式识别与仿真：2007,26(8)：80-83.
[8]李兆烽,李言俊.基于对数极坐标变换的景象匹配新方法.红外与激光工程：2008,37(2)：285-288.
[9]Bailey HH, Blaek well FW. Imgae Correlation. Simulationnad Analysis. AD-AO36665/SSL:1976.
[10]R.O. Tlimieri, M. An, C. Lu. Mhatematies of Multi dimensional. Fourier Tarnsofmr Algorithms Springer:1997.7(18).
[11]Lryna Gordon and David G Lowe. Scene modelling reeognition and traeking with Invariant image features. International Sym Posiumon Mixedand Auglnented Reality(ISMAR):2004,110-119.
[12]Juho Kannala, Esa Rahtu, Janne Heikkila.Amne registration with multi-seale autoconvolution [C]. IEEE International Conference on Image Processing,2005,3:064-1067
[13]Matthew Brownand David GLowe. Automatic Panoramic image stitehing using Invariant features. International Jouma lof Computer Vision,2007,74(1):59-73.
[14]Krystian Mikolajezyk, Cordelia Schmid. Seale & Affine Invariant Interest Point Detectors. International Jouranal of Computer Vison,2004,60(1):63-68.
[15]张凯,抗任意旋转与尺度变化的图像匹配方法研究：(硕十论文).广州：中南大学,2008.
[16]熊辉,基于模板匹配的金刚石颗粒图像识别：(硕十论文).四川：四川大学,2005.
[17]丁慧珍,抗任意角度旋转灰度匹配方法研究：(硕十学位论文).南京：河海大学,2006.
[18]吴国贤,基于CCD技术的平面机构运动研究：(硕十学位论文).天津：天津大学,2006.
[19]陶晓勋,基于特征的快速抗旋转图像匹配方法研究：(硕十学位论文).南京：河海大学,2007.
[20]Leese J A, Novka C S, ClarkBB.An Automated Tecnique for obtaining cloud motion from Gecsynehrous Satellite Daat using Cross2correlation. APPI Meteoorl,1971,10(2):110-132.
[21]Bnarea D I, Silvermna H F A class algorithms for fast digital imag regisrtation. IEEE Trnas:On Compute,1972,21(2):179-186.
[22]哈勤和,张畅.图像的快速一维投影模板匹配.计算机工程与应用,1998,34(1)：36-38.
[23]陈宁江,李介谷.归一化灰度组合法进行图像匹配.红外与激光工程,20002,9(5)：5-9.
[24]罗钟铱,刘成明.灰度图像匹配的快速算法.计算机辅助设计与图形学报,2005,17(5)：966-970.
[25]山海涛,郭建星,耿则勋.影像匹配中几种相似性测度的分析.测绘信息工程,2003,284：11-13.
[26]张建国.不变性纹理分析：(硕十论文).北京：中国科学院自动化所,2002.
[27]刘进.不变量特征的构造及其在目标识别中的应用.(硕十论文).武汉：华中科技大学,2004.
[28]刘方.目标3D不变特征的提取于应用：(硕十论文).长沙：国防科技大学,2001.
[29]杜杰.两种基于灰度的快速图像匹配算法：(硕十论文).大连：大连海事大学,2007.
[30]CHEN Z, HU X D, JIN C Y, etal... Nanoscale measurement technique of in-plane motion for MEMS based on phase correlation method. SPIE,2006,6357:1-7.
[31]Alireza K, Hong Yawhua.Invariant image recognition by Zernike moments. IEEE Trans on PAMI, 1990.12(5):489-497.
[32]Chin R T. On Image Analysis by the Methods of Moments. IEEE Trans.on Pattern Analysis and Machine Intelligence,1988,10(4):496-513.
[33]武文辉.基于Zernike矩和DCT变换的数字水印算法.计算机工程,2009,35(13)：174-176.
[34]杜登崇.可同时估计旋转和平移的电子稳像方法.计算机工程与应用,2010,46(4)：233-235.
[35]陈治.基于相位相关技术的MEMS旋转角度高分辨力测量.光学精密工程,2009,17(8)：1884-1889.
[36]Ullah F,Kaneko S. Using orientation eodes for rotation-invariant template matching. Pattern Recognition,2004,37(2):201-209.
[37]董必立.基于带反馈的图像金字塔的目标的检测与研究：(硕十论文).广州：华南师范大学2008.
[38]王培容,陈鸿雁,李绞军.基于方向码的模板匹配方法.计算机工程与设计,2007,28(14)：3407-3409.
[39]徐旦华,辜嘉,李松毅.Zernike矩的快速算法.东南大学学报,2002,32(2)：189-192.
[40]Tsai Du-Ming, Chiang Cheng-Huei. Rotation-invariant pattern matehing using wavelet Decomposition, Pattern Recognition Letters,2002,23:191-201.
[41]徐亦斌,王敬东,李鹏.基于圆投影向量的景象匹配方法研究.系统工程电子技术,2005,27(10)：1725-1728.
[42]孙君顶,郭启强.基于曲率尺度空间的角点检测.光电工程,2009,36(7)：78-82.
[43]Hu MK. Visual pattern Recognition by Moment Invariants. IREIrans. Information theory, 1962,IT(8):179-187.
[44]潘秀琴,催可宁,侯朝桢.基于局部小波矩的图像匹配算法.计算机工程与应用.2002,6(11)：18-20.
[45]冯桂,卢健,林宗坚.图像直方图不变特征在影像匹配定位中的应用.计算机辅助设计与图形学报,2000,12(2)：146-148.
[46]聂恒,赵荣椿,康宝生.基于边缘几何特征的图像精确匹配方法.计算机辅助设计与图形学报,2004,16(12)：1668-1673.
[47]M. Jogan, etal, Karhunen-Loeve. Expansion of a set of rotated telates. IEEE Trans. Image Professing,2003,12(7):817-825.
[48]罗成平,龚沛曾.图像匹配技术.微型电脑应用,2000,16(3)：26-29.
[49]邵蕊.指针式仪表识别方法的研究：(硕十论文).沈阳：沈阳工业大学,2009.
[50]李媛.基于图像识别的表盘读数系统研究：(硕十学位论文).沈阳：沈阳工业大学,2010.
[51]Xiao, Chen He. Corner detector based on global and local curvature properties. Optical Engineering. (SPIE) 2008,47(5):465-469.
[52]S. M. Simth, J. M. Brady. SUSAN-A New Approach to Low Level Image Processing. Defense Research Agency,1994,I-JCV(23):45-48.
[53]A. Rattarangsi, R. T. Chin. Scale-based detection of corners of planar Curves. IEEE Trans on Pattern Analysis and Ma-chine Intelligence,1992,14(4):430-449.
[54]C. Urdiales, C. Trazegnies, A. Bandera, and F. Sandoval. Corner detection based on an adaptively filtered curvature function, Electron. Lett.2003,39:426-428.
[55]A. Rattarangsi, R. T. Chin. "Scale-based detection of corners of planar Curves"[J]. IEEE Trans on Pattern Analysis and Ma-chine Intelligence,1992,14(4):430-449.
[56]汪华琴,谈国新.一种基于曲率尺度空间的自适应角点检测算法.计算机技术与自动化,2007,26(2)：145-148.
[57]陈宏.基于形状特征的图像检索：(硕士论文).西安：西安电子科技大学,2010.
[58]张亮.号码技术在若干领域应用的研究：(硕十论文).沈阳：沈阳工业大学,2006.
[59]Ullah, F,Kaneko. S.Using orientation codes for rotation invariant templete matching. Pattern Recognition,2004,37(2):201-209.
[60]郭丽,孙兴华.基于子块均值的快速匹配算法.计算机应用,2006,26(7)：1595-1597.
[61]Farhan Ullah, Shunichi Kaneko. Using orientation codes for rota-tion-invariant template. Pattern Recognition:2004.
[62]WANG Jue, SUN Xiao-wei. Image Matching Based on the Invariant Feature of Circle rojection and Histogram. Pattern Recognition and Simuation:2007,26(8):80-83.
[63]王立,李言俊.对数极坐标变换识别算法在成像制导中的应用.宇航学报,2005,26(3).
[64]刘惠中,李言俊.空间变分辨率机理在红外成像制导中的应用研究.弹道与制导学报,2005,25(2).
[2]Ehlers.M, Fuller M. Region-based matching for image integration in remote sensing databases. Dig Int Geosci mote Sens Symp (IGARSS):1991(4):2231-2234.
[3]Brown L G.. Survey of image registration techniques. ACM Computing Surveys:1992,24(4): 325-376.
[4]Sim D G, Kim H K, etal. Translation scale, and rotation in variant texture descriptor for texture-based image retrieval. IEEE Int ConfImage Process:2000,3:742-745.
[5]Farhan Ullah, Shunichi Kaneko. Using orientation codes for rota-tion-invariant template matching. Pattern Recognition:2004.
[6]Sang-Woo Lee, Whoi-Yul. Kim. Rotation-invariant template matching using projection method. Proc KITE:1996,19(1):475-476.
[7]王钰,孙小惟.基于圆投影及直方图不变特征的图像匹配方法.模式识别与仿真：2007,26(8)：80-83.
[8]李兆烽,李言俊.基于对数极坐标变换的景象匹配新方法.红外与激光工程：2008,37(2)：285-288.
[9]Bailey HH, Blaek well FW. Imgae Correlation. Simulationnad Analysis. AD-AO36665/SSL:1976.
[10]R.O. Tlimieri, M. An, C. Lu. Mhatematies of Multi dimensional. Fourier Tarnsofmr Algorithms Springer:1997.7(18).
[11]Lryna Gordon and David G Lowe. Scene modelling reeognition and traeking with Invariant image features. International Sym Posiumon Mixedand Auglnented Reality(ISMAR):2004,110-119.
[12]Juho Kannala, Esa Rahtu, Janne Heikkila.Amne registration with multi-seale autoconvolution [C]. IEEE International Conference on Image Processing,2005,3:064-1067
[13]Matthew Brownand David GLowe. Automatic Panoramic image stitehing using Invariant features. International Jouma lof Computer Vision,2007,74(1):59-73.
[14]Krystian Mikolajezyk, Cordelia Schmid. Seale & Affine Invariant Interest Point Detectors. International Jouranal of Computer Vison,2004,60(1):63-68.
[15]张凯,抗任意旋转与尺度变化的图像匹配方法研究：(硕十论文).广州：中南大学,2008.
[16]熊辉,基于模板匹配的金刚石颗粒图像识别：(硕十论文).四川：四川大学,2005.
[17]丁慧珍,抗任意角度旋转灰度匹配方法研究：(硕十学位论文).南京：河海大学,2006.
[18]吴国贤,基于CCD技术的平面机构运动研究：(硕十学位论文).天津：天津大学,2006.
[19]陶晓勋,基于特征的快速抗旋转图像匹配方法研究：(硕十学位论文).南京：河海大学,2007.
[20]Leese J A, Novka C S, ClarkBB.An Automated Tecnique for obtaining cloud motion from Gecsynehrous Satellite Daat using Cross2correlation. APPI Meteoorl,1971,10(2):110-132.
[21]Bnarea D I, Silvermna H F A class algorithms for fast digital imag regisrtation. IEEE Trnas:On Compute,1972,21(2):179-186.
[22]哈勤和,张畅.图像的快速一维投影模板匹配.计算机工程与应用,1998,34(1)：36-38.
[23]陈宁江,李介谷.归一化灰度组合法进行图像匹配.红外与激光工程,20002,9(5)：5-9.
[24]罗钟铱,刘成明.灰度图像匹配的快速算法.计算机辅助设计与图形学报,2005,17(5)：966-970.
[25]山海涛,郭建星,耿则勋.影像匹配中几种相似性测度的分析.测绘信息工程,2003,284：11-13.
[26]张建国.不变性纹理分析：(硕十论文).北京：中国科学院自动化所,2002.
[27]刘进.不变量特征的构造及其在目标识别中的应用.(硕十论文).武汉：华中科技大学,2004.
[28]刘方.目标3D不变特征的提取于应用：(硕十论文).长沙：国防科技大学,2001.
[29]杜杰.两种基于灰度的快速图像匹配算法：(硕十论文).大连：大连海事大学,2007.
[30]CHEN Z, HU X D, JIN C Y, etal... Nanoscale measurement technique of in-plane motion for MEMS based on phase correlation method. SPIE,2006,6357:1-7.
[31]Alireza K, Hong Yawhua.Invariant image recognition by Zernike moments. IEEE Trans on PAMI, 1990.12(5):489-497.
[32]Chin R T. On Image Analysis by the Methods of Moments. IEEE Trans.on Pattern Analysis and Machine Intelligence,1988,10(4):496-513.
[33]武文辉.基于Zernike矩和DCT变换的数字水印算法.计算机工程,2009,35(13)：174-176.
[34]杜登崇.可同时估计旋转和平移的电子稳像方法.计算机工程与应用,2010,46(4)：233-235.
[35]陈治.基于相位相关技术的MEMS旋转角度高分辨力测量.光学精密工程,2009,17(8)：1884-1889.
[36]Ullah F,Kaneko S. Using orientation eodes for rotation-invariant template matching. Pattern Recognition,2004,37(2):201-209.
[37]董必立.基于带反馈的图像金字塔的目标的检测与研究：(硕十论文).广州：华南师范大学2008.
[38]王培容,陈鸿雁,李绞军.基于方向码的模板匹配方法.计算机工程与设计,2007,28(14)：3407-3409.
[39]徐旦华,辜嘉,李松毅.Zernike矩的快速算法.东南大学学报,2002,32(2)：189-192.
[40]Tsai Du-Ming, Chiang Cheng-Huei. Rotation-invariant pattern matehing using wavelet Decomposition, Pattern Recognition Letters,2002,23:191-201.
[41]徐亦斌,王敬东,李鹏.基于圆投影向量的景象匹配方法研究.系统工程电子技术,2005,27(10)：1725-1728.
[42]孙君顶,郭启强.基于曲率尺度空间的角点检测.光电工程,2009,36(7)：78-82.
[43]Hu MK. Visual pattern Recognition by Moment Invariants. IREIrans. Information theory, 1962,IT(8):179-187.
[44]潘秀琴,催可宁,侯朝桢.基于局部小波矩的图像匹配算法.计算机工程与应用.2002,6(11)：18-20.
[45]冯桂,卢健,林宗坚.图像直方图不变特征在影像匹配定位中的应用.计算机辅助设计与图形学报,2000,12(2)：146-148.
[46]聂恒,赵荣椿,康宝生.基于边缘几何特征的图像精确匹配方法.计算机辅助设计与图形学报,2004,16(12)：1668-1673.
[47]M. Jogan, etal, Karhunen-Loeve. Expansion of a set of rotated telates. IEEE Trans. Image Professing,2003,12(7):817-825.
[48]罗成平,龚沛曾.图像匹配技术.微型电脑应用,2000,16(3)：26-29.
[49]邵蕊.指针式仪表识别方法的研究：(硕十论文).沈阳：沈阳工业大学,2009.
[50]李媛.基于图像识别的表盘读数系统研究：(硕十学位论文).沈阳：沈阳工业大学,2010.
[51]Xiao, Chen He. Corner detector based on global and local curvature properties. Optical Engineering. (SPIE) 2008,47(5):465-469.
[52]S. M. Simth, J. M. Brady. SUSAN-A New Approach to Low Level Image Processing. Defense Research Agency,1994,I-JCV(23):45-48.
[53]A. Rattarangsi, R. T. Chin. Scale-based detection of corners of planar Curves. IEEE Trans on Pattern Analysis and Ma-chine Intelligence,1992,14(4):430-449.
[54]C. Urdiales, C. Trazegnies, A. Bandera, and F. Sandoval. Corner detection based on an adaptively filtered curvature function, Electron. Lett.2003,39:426-428.
[55]A. Rattarangsi, R. T. Chin. "Scale-based detection of corners of planar Curves"[J]. IEEE Trans on Pattern Analysis and Ma-chine Intelligence,1992,14(4):430-449.
[56]汪华琴,谈国新.一种基于曲率尺度空间的自适应角点检测算法.计算机技术与自动化,2007,26(2)：145-148.
[57]陈宏.基于形状特征的图像检索：(硕士论文).西安：西安电子科技大学,2010.
[58]张亮.号码技术在若干领域应用的研究：(硕十论文).沈阳：沈阳工业大学,2006.
[59]Ullah, F,Kaneko. S.Using orientation codes for rotation invariant templete matching. Pattern Recognition,2004,37(2):201-209.
[60]郭丽,孙兴华.基于子块均值的快速匹配算法.计算机应用,2006,26(7)：1595-1597.
[61]Farhan Ullah, Shunichi Kaneko. Using orientation codes for rota-tion-invariant template. Pattern Recognition:2004.
[62]WANG Jue, SUN Xiao-wei. Image Matching Based on the Invariant Feature of Circle rojection and Histogram. Pattern Recognition and Simuation:2007,26(8):80-83.
[63]王立,李言俊.对数极坐标变换识别算法在成像制导中的应用.宇航学报,2005,26(3).
[64]刘惠中,李言俊.空间变分辨率机理在红外成像制导中的应用研究.弹道与制导学报,2005,25(2).