珍珠图像的预处理与特征检测
|
摘要
珍珠产业是中国的传统产业和民族产业,我国是世界第一的珍珠大国。目前,珍珠的等级评判主要靠人工依据珍珠的颜色、光泽、形状、纹理等进行大致估计。这种估计在很大程度上受人为因素的影响,导致人为误差,造成分选质量差、人工成本高、分选效率低、随意性大,从而给整个珍珠生产业带来较大的负面影响。因此根据珍珠的特点,建立详细的分级标准、研制高速、高精度珍珠自动分选设备对我国珍珠业的发展将起到重要的促进作用。
珍珠自动分选系统包括硬件系统、软件系统。珍珠自动分选系统的硬件部分,主要包括照明系统,触发设备,工业相机,IO控制卡,传送设备和分类器几个部分。软件系统主要由边缘检测、形状检测、纹理检测、颜色识别等部分构成。珍珠自动分选设备通过对珍珠进行动态采集,抓取珍珠的图象,对所抓取图象进行分析处理,从而按照珍珠的大小、形状、瑕疵、颜色、光泽等特征指标进行分类。
珍珠的形状主要有圆形、椭圆形、扁形、其他形状等。本系统利用随机Hough变换技术来检测珍珠形状,能够检测圆、椭圆、尖形和其它形状。珍珠纹理以珍珠表面的瑕疵情况作为分选标准,有无暇、微暇、小暇、瑕疵、重暇等。珍珠图像经边缘检测找到边界轮廓后,再对图像进行二值化处理,提取瑕疵信息,然后进行瑕疵识别,并确定瑕疵类型。
Pearl industry is China's traditional and national industry, China is the world's major countries of pearls. At present, the pearl level judging depends largely on the artificial estimate of the color, luster, shape, texture of the pearl. This estimate in large measure by man-made factors, lead to human error, resulting in poor quality, high labor costs, low efficiency of separation, arbitrariness, and thus to the entire pearl production industry greater negative impact.
Pearl automatic sorting system, include hardware, software systems. The hardware system includes lighting systems, triggering equipment, industrial cameras, IO control card, and parts for classification. And software system mainly by the edge detection, shape detection, texture detection, and color detection. In our system, by dynamic collecting the pearl image, we define the class of pearls according to the shape, color, luster and flaw of pearls.
Mainly the shape of pearls round, oval, flat, the other shape, and so on. In this paper, we use Hough transformation to detect the shape of pearls. It can recognize round, ellipse and sharp shape. By the image edge detection, we find the border of the pearl, then we binary the pearl image, and fetch the texture information and proceed to identify flaws and defects type.
珍珠自动分选系统包括硬件系统、软件系统。珍珠自动分选系统的硬件部分,主要包括照明系统,触发设备,工业相机,IO控制卡,传送设备和分类器几个部分。软件系统主要由边缘检测、形状检测、纹理检测、颜色识别等部分构成。珍珠自动分选设备通过对珍珠进行动态采集,抓取珍珠的图象,对所抓取图象进行分析处理,从而按照珍珠的大小、形状、瑕疵、颜色、光泽等特征指标进行分类。
珍珠的形状主要有圆形、椭圆形、扁形、其他形状等。本系统利用随机Hough变换技术来检测珍珠形状,能够检测圆、椭圆、尖形和其它形状。珍珠纹理以珍珠表面的瑕疵情况作为分选标准,有无暇、微暇、小暇、瑕疵、重暇等。珍珠图像经边缘检测找到边界轮廓后,再对图像进行二值化处理,提取瑕疵信息,然后进行瑕疵识别,并确定瑕疵类型。
Pearl industry is China's traditional and national industry, China is the world's major countries of pearls. At present, the pearl level judging depends largely on the artificial estimate of the color, luster, shape, texture of the pearl. This estimate in large measure by man-made factors, lead to human error, resulting in poor quality, high labor costs, low efficiency of separation, arbitrariness, and thus to the entire pearl production industry greater negative impact.
Pearl automatic sorting system, include hardware, software systems. The hardware system includes lighting systems, triggering equipment, industrial cameras, IO control card, and parts for classification. And software system mainly by the edge detection, shape detection, texture detection, and color detection. In our system, by dynamic collecting the pearl image, we define the class of pearls according to the shape, color, luster and flaw of pearls.
Mainly the shape of pearls round, oval, flat, the other shape, and so on. In this paper, we use Hough transformation to detect the shape of pearls. It can recognize round, ellipse and sharp shape. By the image edge detection, we find the border of the pearl, then we binary the pearl image, and fetch the texture information and proceed to identify flaws and defects type.
引文
[1]M.Recce,J.Taylor,A.Plebe,G.Tropiano.Vision and neural control for an orange harvesting robot[C].Proceedings of the 1996 International Workshop on Neural Networks for Identification,Control,Robotics,and Signal/Image Processing,1996:467-475
[2]P.H.Heinemann,N.P.Pathare,C.T.Morrow.An automated inspection station for machine-vision grading of potatoes[J].Machine Vision and Applications,1996,9(1):14-19
[3]D.M.Bulanon,T.Kataoka,S.Zhang,et al.Optimal thresholding for the automatic recognition of apple fruits[R].An ASAE Meeting Presentation,2001
[4]B.Li,G.Li,Y.Wang,et al.A classification method of pearl shape based on Zernike moment[C].Proceedings of the 2007 International Conference on Wavelet Analysis and Pattern Recognition,2007,3:1076-1079
[5]J.Zhao,D.He.Research of computer recognition method on fruit shape[J].Journal of Agricultural Engineering,2001,17(2):165-167
[6]S.Guan,H.Zhou,G.Liu.Fuzzy classification of rice kernel shape based on Zernike moments[J].Henan University of Technology Journal,2005,26(1):23-25
[7]崔屹.数字图像处理技术与应用[M].北京电子工业出版社,1997:56-1601
[8]田岩,田金文,柳健等.超分辨率技术的实现-一种改善的小波插值方法[J].中国图像图形学报,2003,8(12):1422-1425
[9]吕俊哲.图像二值化算法研究及其实现[J].科技情报开发与经济,2004,14(12):266-267
[10]张雅兰.图像的二值化处理[J].广西工学院学报,2002,13(1):33-55
[11]吴一全,朱兆达.图像处理中阈值选取方法30年(1962-1992)的进展(一)[J].数据采集与处理,1993,8(3):193-201.
[12]N.Ostu.A threshold selection method from gray-level histograms[J].IEEE Transactions on SMC,1979,9(2):62-66
[13]T.Poggio,H.Voorhees,A.Yuille.A regularized solution to edge detection[R].Journal of Complexity,1988,4(2):106-123
[14]魏伟波,芮筱亭.图像边缘检测方法研究[J].计算机工程与应用,2006,42(30):88-91
[15]范江艳,史金飞,张磊.几种边缘检测算子在半导体器件视觉检测中的应用分析[J].电器技术与自动化,2006,35(5):130-132
[16]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,2006,16(14):53-56
[17]J.Canny.A computational approach to edge detection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1986,8(6):679-698
[18]M.Unser,A.Aldroubi,M.Eden.On the asymptotic convergence of B-spline wavelets to Gabor functions[J].IEEE Transactions on Information Theory,1992,38(2):864-872
[19]刘曙光,刘明远,何钺.基于Canny准则的基数B样条小波边缘检测[J].信号处理,2001,17(5):418-423
[20]J.Illingworth,J.Kittler.A survey of the Hough transform[J].Computer Vision Graphics Image Process,1988,44(1):87-116
[21]L.Xu,E.Oja.Randomized Hough transform(RHT):basie mechanisms,algorithms and computational complexities[J].Computer Vision Graphic Image Process:Image understanding,1993,57(2):131-154
[22]A.R.Hare,M.B.Sandier.Improved-performance randomised Hough transform.Electronics letters,1992,28(18):1678-1680
[23]T.Chen,K.Chung.An efficient randomized algorithm for detecting circles.Computer Vision and Image Understanding,2001,83(2):172-191
[24]R.O.Duda,P.E.Hart.Use of the Hough transformation to detect lines and curves in pictures.Communications of the ACM,1972,15(1):11-15
[25]X.Cao,F.Deravi.An efficient method for multiple-circle detection.Computer Vision,1990.Proceedings,Third International Conference on,1990:744-747
[26]陈燕新,戚飞虎.基于随机Hough变换的快速圆检测方法.上海交通大学学 报,1998,32(10):17-20
[27]齐晓娟,董明利,王君等.数字摄影测量中椭圆的高效检测方法.北京机械工业学院学报,2005,20(4):5-7
[28]庞永杰,刘维.改进的K-RANSAC椭圆检测方法[J].哈尔滨工程大学学报,2003,24(5):487-490
[29]陈燕新,戚飞虎.一种新的基于随机Hough变换的椭圆检测方法.红外与毫米波学报,2000,19(1):43-47
[30]陈燕新,戚飞虎.利用梯度方向信息的随机Hough变换.红外与毫米波学报,1998,17(5):375-380
[31]L.S.Davis.Image teture analysis techniques - a survey.Digital Image Processing,1981:189-201
[32]张鹤等.基于灰度共生矩阵的图像纹理分析.中国科技论文在线,2004
[33]白雪冰,王克奇,王辉.基于灰度共生矩阵的木材纹理分类方法的研究(J).哈尔滨工程大学学报,2005,37(12):1667-1670
[2]P.H.Heinemann,N.P.Pathare,C.T.Morrow.An automated inspection station for machine-vision grading of potatoes[J].Machine Vision and Applications,1996,9(1):14-19
[3]D.M.Bulanon,T.Kataoka,S.Zhang,et al.Optimal thresholding for the automatic recognition of apple fruits[R].An ASAE Meeting Presentation,2001
[4]B.Li,G.Li,Y.Wang,et al.A classification method of pearl shape based on Zernike moment[C].Proceedings of the 2007 International Conference on Wavelet Analysis and Pattern Recognition,2007,3:1076-1079
[5]J.Zhao,D.He.Research of computer recognition method on fruit shape[J].Journal of Agricultural Engineering,2001,17(2):165-167
[6]S.Guan,H.Zhou,G.Liu.Fuzzy classification of rice kernel shape based on Zernike moments[J].Henan University of Technology Journal,2005,26(1):23-25
[7]崔屹.数字图像处理技术与应用[M].北京电子工业出版社,1997:56-1601
[8]田岩,田金文,柳健等.超分辨率技术的实现-一种改善的小波插值方法[J].中国图像图形学报,2003,8(12):1422-1425
[9]吕俊哲.图像二值化算法研究及其实现[J].科技情报开发与经济,2004,14(12):266-267
[10]张雅兰.图像的二值化处理[J].广西工学院学报,2002,13(1):33-55
[11]吴一全,朱兆达.图像处理中阈值选取方法30年(1962-1992)的进展(一)[J].数据采集与处理,1993,8(3):193-201.
[12]N.Ostu.A threshold selection method from gray-level histograms[J].IEEE Transactions on SMC,1979,9(2):62-66
[13]T.Poggio,H.Voorhees,A.Yuille.A regularized solution to edge detection[R].Journal of Complexity,1988,4(2):106-123
[14]魏伟波,芮筱亭.图像边缘检测方法研究[J].计算机工程与应用,2006,42(30):88-91
[15]范江艳,史金飞,张磊.几种边缘检测算子在半导体器件视觉检测中的应用分析[J].电器技术与自动化,2006,35(5):130-132
[16]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,2006,16(14):53-56
[17]J.Canny.A computational approach to edge detection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1986,8(6):679-698
[18]M.Unser,A.Aldroubi,M.Eden.On the asymptotic convergence of B-spline wavelets to Gabor functions[J].IEEE Transactions on Information Theory,1992,38(2):864-872
[19]刘曙光,刘明远,何钺.基于Canny准则的基数B样条小波边缘检测[J].信号处理,2001,17(5):418-423
[20]J.Illingworth,J.Kittler.A survey of the Hough transform[J].Computer Vision Graphics Image Process,1988,44(1):87-116
[21]L.Xu,E.Oja.Randomized Hough transform(RHT):basie mechanisms,algorithms and computational complexities[J].Computer Vision Graphic Image Process:Image understanding,1993,57(2):131-154
[22]A.R.Hare,M.B.Sandier.Improved-performance randomised Hough transform.Electronics letters,1992,28(18):1678-1680
[23]T.Chen,K.Chung.An efficient randomized algorithm for detecting circles.Computer Vision and Image Understanding,2001,83(2):172-191
[24]R.O.Duda,P.E.Hart.Use of the Hough transformation to detect lines and curves in pictures.Communications of the ACM,1972,15(1):11-15
[25]X.Cao,F.Deravi.An efficient method for multiple-circle detection.Computer Vision,1990.Proceedings,Third International Conference on,1990:744-747
[26]陈燕新,戚飞虎.基于随机Hough变换的快速圆检测方法.上海交通大学学 报,1998,32(10):17-20
[27]齐晓娟,董明利,王君等.数字摄影测量中椭圆的高效检测方法.北京机械工业学院学报,2005,20(4):5-7
[28]庞永杰,刘维.改进的K-RANSAC椭圆检测方法[J].哈尔滨工程大学学报,2003,24(5):487-490
[29]陈燕新,戚飞虎.一种新的基于随机Hough变换的椭圆检测方法.红外与毫米波学报,2000,19(1):43-47
[30]陈燕新,戚飞虎.利用梯度方向信息的随机Hough变换.红外与毫米波学报,1998,17(5):375-380
[31]L.S.Davis.Image teture analysis techniques - a survey.Digital Image Processing,1981:189-201
[32]张鹤等.基于灰度共生矩阵的图像纹理分析.中国科技论文在线,2004
[33]白雪冰,王克奇,王辉.基于灰度共生矩阵的木材纹理分类方法的研究(J).哈尔滨工程大学学报,2005,37(12):1667-1670