一种包装卷材快速在线检测算法
|
摘要
目的提高包装卷材图像的检测速度与精度,以实现快速在线检测。方法预先对几副标准样本图像进行拼接,将得到拓展的样板图像作为动态模板的母版。检测过程中,首先计算待测图样中心区域经简化的图像几何矩,并在前述样板中的一段区间内搜寻最佳匹配位置。再以该位置为中心,在样板图像上划分出与待测图样相同大小的区域作为动态模板,最后利用图像差影法实现缺陷检测与识别。结果二阶简化几何矩能有效减少运算量,动态模板可显著提高卷材图像检测速度,以电子元器件包装卷材载料带为检测对象对所提出的方法进行了测试,在较高分辨率条件下检测速度不低于2 m/s。结论该算法可满足包装卷材快速在线检测需求,对提高相位不确定性图像检测速度有一定参考意义。
The work aims to improve the detection speed and accuracy of the packaging coil image to achieve the rapid online detection. A few standard image samples were spliced beforehand. The extended image samples obtained were taken as the master mask of the variable template. A simplified image geometric moment in the central region of the image sample to be detected should be calculated first during the detection, and the best matching position was searched in a section of the foregoing template. Then, with such position as the center, the area marked off on the template image having the same size as the image to be detected was regarded as the variable template. Finally, the image subtraction method was implemented to realize the detection and recognition of defects. The two-order simplified geometric moments could reduce the operand effectively, and the dynamic template could significantly improve the detection speed of the coil image. The proposed algorithm was tested with the packaging coil carrier tape of the electronic components as the detected objects. The detection speed was no less than 2 m/s under the condition of higher resolution. The proposed algorithm can meet the requirements of rapid online detection of packaging coil, and provide certain reference significance for the detection speed of uncertain-phase image.
The work aims to improve the detection speed and accuracy of the packaging coil image to achieve the rapid online detection. A few standard image samples were spliced beforehand. The extended image samples obtained were taken as the master mask of the variable template. A simplified image geometric moment in the central region of the image sample to be detected should be calculated first during the detection, and the best matching position was searched in a section of the foregoing template. Then, with such position as the center, the area marked off on the template image having the same size as the image to be detected was regarded as the variable template. Finally, the image subtraction method was implemented to realize the detection and recognition of defects. The two-order simplified geometric moments could reduce the operand effectively, and the dynamic template could significantly improve the detection speed of the coil image. The proposed algorithm was tested with the packaging coil carrier tape of the electronic components as the detected objects. The detection speed was no less than 2 m/s under the condition of higher resolution. The proposed algorithm can meet the requirements of rapid online detection of packaging coil, and provide certain reference significance for the detection speed of uncertain-phase image.
引文
[1]JIANG Y,MA S,GAO H,et al.Research on Defect Detection Technology of Tablets in Aluminum Plastic Package[J].Open Automation&Control Systems Journal,2014,6(1):940—951.
[2]舒文娉,刘全香.基于支持向量机的印品缺陷分类方法[J].包装工程,2014,35(23):138—142.SHU Wen-ping,LIU Quan-xiang.Classification Method of Printing Defects Based on Support Vector Machine[J].Packaging Engineering,2014,35(23):138—142.
[3]黄杰贤,李迪,黄志平,等.FPC焊盘表面缺陷检测研究[J].激光与红外,2014(6):692—696.HUANG Jie-xian,LI Di,HUANG Zhi-ping,et al.Study on Defect Detection of Flexible Printed Circuit Solder Surface[J].Laser and Infrared,2014(6):692—696.
[4]SCHWEITZER H,RUI D,ANDERSON R F.A Dual-Bound Algorithm for Very Fast and Exact Template Matching[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2011,33(3):459—470.
[5]LI Y,LI H,CAI Z.Fast Orthogonal Haar Transform Pattern Matching via Image Square Sum[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2014,36(9):1748—1760.
[6]苑玮琦,李绍丽,李德健.基于子区域变尺度高斯拟合的木材表面缺陷识别[J].仪器仪表学报,2016,37(4):879—886.YUAN Wei-qi,LI Shao-li,LI De-jian.Wood Surface Defect Recognition Based on Sub-region Zoom Gaussian Fitting[J].Chinese Journal of Scientific Instrument,2016,37(4):879—886.
[7]陈丽,唐万有.基于DP方法的印刷品图像检测技术研究[J].包装工程,2014,35(5):116—120.CHEN Li,TANG Wan-you.Image Detection Technology of Printing Based on the DP Method[J].Packaging Engineering,2014,35(5):116—120.
[8]张秀珍,吴贵芳,普杰信,等.基于模板匹配的彩色印品套印偏差检测方法[J].计算机仿真,2015,32(10):250—253.ZHANG Xiu-zhen,WU Gui-fan,PU Jie-xin,et al.A Method for Color Overprint Deviation Inspect Based on Template Matching[J].Computer Simulation,2015,32(10):250—253.
[9]陈皓,马彩文,陈岳承,等.基于灰度统计的快速模板匹配算法[J].光子学报,2009,38(6):1586—1590.CHEN Hao,MA Cai-wen,CHEN Yue-cheng,et al.Fast Template Matching Algorithm Based on Gray Statistics[J].Acta Photonica Sinica,2009,38(6):1586—1590.
[10]RABINER L.On the Use of Autocorrelation Analysis for Pitch Detection[J].IEEE Transactions on Acoustics Speech&Signal Processing,2003,25(1):24—33.
[11]SONDHI M M.New Methods of Pitch Extraction[J].IEEE Transactions on Audio&Electroacoustics,1968,16(2):262—266.
[12]马强,项昭保,黄良学,等.基于改进SIFT和RANSAC图像拼接算法研究[J].计算机技术与发展,2016,26(4):61—65.MA Qiang,XIANG Zhao-bao,HUANG Liang-xue,et al.Research on Panorama Image Mosaic Algorithm Based on Improved SIFT and RANSAC[J].Computer Technology and Development,2016,26(4):61—65.
[13]LIU W,RIBEIRO E.Incremental Variations of Image Moments for Nonlinear Image Registration[J].Signal Image&Video Processing,2014,8(3):423—432.
[14]LIU W,RIBEIRO E.Nonrigid Image Registration Using Moments[M].Emerging Topics in Computer Vision and Its Applications,2014.
[15]ZHAO Y M,XIE W F,LIU S,et al.Neural NetworkBased Image Moments for Robotic Visual Servoing[J].Journal of Intelligent&Robotic Systems,2015,78(2):239—256.
[16]张朝鑫,席平.高斯几何矩及其在特征匹配与图像配准中的应用[J].计算机辅助设计与图形学学报,2014,26(7):1116—1125.ZHANG Chao-xin,XI Ping.Gaussian-Geometric Moments and Its Application in Feature Matching&Image Registration,Journal of Computer-aided Design and Computer Graphics,2014,26(7):1116—1125.
[2]舒文娉,刘全香.基于支持向量机的印品缺陷分类方法[J].包装工程,2014,35(23):138—142.SHU Wen-ping,LIU Quan-xiang.Classification Method of Printing Defects Based on Support Vector Machine[J].Packaging Engineering,2014,35(23):138—142.
[3]黄杰贤,李迪,黄志平,等.FPC焊盘表面缺陷检测研究[J].激光与红外,2014(6):692—696.HUANG Jie-xian,LI Di,HUANG Zhi-ping,et al.Study on Defect Detection of Flexible Printed Circuit Solder Surface[J].Laser and Infrared,2014(6):692—696.
[4]SCHWEITZER H,RUI D,ANDERSON R F.A Dual-Bound Algorithm for Very Fast and Exact Template Matching[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2011,33(3):459—470.
[5]LI Y,LI H,CAI Z.Fast Orthogonal Haar Transform Pattern Matching via Image Square Sum[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2014,36(9):1748—1760.
[6]苑玮琦,李绍丽,李德健.基于子区域变尺度高斯拟合的木材表面缺陷识别[J].仪器仪表学报,2016,37(4):879—886.YUAN Wei-qi,LI Shao-li,LI De-jian.Wood Surface Defect Recognition Based on Sub-region Zoom Gaussian Fitting[J].Chinese Journal of Scientific Instrument,2016,37(4):879—886.
[7]陈丽,唐万有.基于DP方法的印刷品图像检测技术研究[J].包装工程,2014,35(5):116—120.CHEN Li,TANG Wan-you.Image Detection Technology of Printing Based on the DP Method[J].Packaging Engineering,2014,35(5):116—120.
[8]张秀珍,吴贵芳,普杰信,等.基于模板匹配的彩色印品套印偏差检测方法[J].计算机仿真,2015,32(10):250—253.ZHANG Xiu-zhen,WU Gui-fan,PU Jie-xin,et al.A Method for Color Overprint Deviation Inspect Based on Template Matching[J].Computer Simulation,2015,32(10):250—253.
[9]陈皓,马彩文,陈岳承,等.基于灰度统计的快速模板匹配算法[J].光子学报,2009,38(6):1586—1590.CHEN Hao,MA Cai-wen,CHEN Yue-cheng,et al.Fast Template Matching Algorithm Based on Gray Statistics[J].Acta Photonica Sinica,2009,38(6):1586—1590.
[10]RABINER L.On the Use of Autocorrelation Analysis for Pitch Detection[J].IEEE Transactions on Acoustics Speech&Signal Processing,2003,25(1):24—33.
[11]SONDHI M M.New Methods of Pitch Extraction[J].IEEE Transactions on Audio&Electroacoustics,1968,16(2):262—266.
[12]马强,项昭保,黄良学,等.基于改进SIFT和RANSAC图像拼接算法研究[J].计算机技术与发展,2016,26(4):61—65.MA Qiang,XIANG Zhao-bao,HUANG Liang-xue,et al.Research on Panorama Image Mosaic Algorithm Based on Improved SIFT and RANSAC[J].Computer Technology and Development,2016,26(4):61—65.
[13]LIU W,RIBEIRO E.Incremental Variations of Image Moments for Nonlinear Image Registration[J].Signal Image&Video Processing,2014,8(3):423—432.
[14]LIU W,RIBEIRO E.Nonrigid Image Registration Using Moments[M].Emerging Topics in Computer Vision and Its Applications,2014.
[15]ZHAO Y M,XIE W F,LIU S,et al.Neural NetworkBased Image Moments for Robotic Visual Servoing[J].Journal of Intelligent&Robotic Systems,2015,78(2):239—256.
[16]张朝鑫,席平.高斯几何矩及其在特征匹配与图像配准中的应用[J].计算机辅助设计与图形学学报,2014,26(7):1116—1125.ZHANG Chao-xin,XI Ping.Gaussian-Geometric Moments and Its Application in Feature Matching&Image Registration,Journal of Computer-aided Design and Computer Graphics,2014,26(7):1116—1125.