物体3D模型恢复中条纹图像的处理
|
摘要
基于光学非接触式测来能够的物体3D模型恢复技术是光学、图像与视觉检测的一个重要分支,该技术广泛应用于反求工程、CAD/CAM设计、模具设计、医疗诊断、在线检测、个性化制造等领域。面结构光投影法能快速、准确地获取三维物体的模型信息,而且测量过程完全非接触、测量速度快,测量数据空间分辨率高和系统成本低,是目前光学非接触式测量的研究热点。条纹图像处理是其关键技术之一。本文对现有的条纹图像处理技术进行了比较分析,对投影条纹的模式及相应的条纹图像处理进行了深入的研究探讨。
本文研究了三基色条纹模板下条纹图像处理及条纹编码方法。三基色条纹模板利用颜色信息提高了光学三维测量的抗干扰能力,有利于条纹断裂处中心线对应条纹的寻找。提出了在该模板下的条纹图像处理算法中无需任何偏色图像先验信息或理想假设条件的图像偏色调整方法;改进了条件细化算法;研究了基于搜索线的条纹自动编码算法。实验表明得到的条纹中心线单像素程度高、居中性好、无扭曲变形、断点少;条纹自动编码结果正确,无漏编、错编现象,处理结果达到工程实际应用的要求。
本文研究了灰度正弦变化条纹模板下条纹图像处理及条纹编码方法。在灰度正弦变化条纹模板中以彩色中央条纹为标志条纹,确立条纹编码基准。利用旋滤波对条纹图像进行预处理,运用二维导数符号二值图法提取条纹中心线,并将所提取的亮条纹中心线与暗条纹中心线赋予不同颜色,以便后续处理中利用色彩的相间规律对条纹进行编码。
本文在分析三基色条纹模板和灰度正弦变化条纹模板各自特点的基础上,研究了三基色灰度正弦变化条纹投射模式及其条纹图像处理方法。
最后,对三种条纹模板下条纹图像处理及条纹编码的实验结果进行了对比分析,总结了三种条纹模板各自的优缺点和适用范围。
Non-contact 3-Dimensiongal optical model reconstruction is an important branch of modern optics, laser technique, digital image processing technique and machine vision, which is widely applied in the fields of reverse engineering, CAD/CAM, mold design, medical treatment, measurement online, individuation manufacturing and so on. The method called face-structure grating projection became a study hot point of 3-Dimensiongal optical model reconstruction. This method can measure the surface of 3D object rapidly and nicely. Further more, it has some advantages, such as non-contract measured process, high measure speed, high resolving ability of data space, little cost and so on. One of its key techniques is the digital stripe image processing. Based on comparing and analyzing the existing techniques, this paper deeply studied the stripe pattern of projecting grating.
The methods of stripe image procession and stripe order recognition in tricolor-coded stripe pattern are studied and proposed, and theoretical analysis and experimental results are given. Tricolor-coded stripe pattern has the advantages of providing information for stripe order recognition, and thereby carries the measurement system high ability of disturbance-resistance. An algorithm is proposed to automatically correct the color of any image without using any additional knowledge or any assumed ideal condition. This paper presents a stripe thinning algorithm by improving Condition Parallel Algorithm. An automatic stripe order recognition and storage method for tricolor-coded stripe projection pattern is also presented in the paper. Finally, the results of the experiments are also presented to prove the effectiveness of the above discussed methods.
The methods of stripe image procession and stripe order recognition in sine gray-value distribution stripe pattern are studied and proposed. Furthermore, theoretical analysis and experimental results are given. With the stripe color ,to determine the symbol stripe can be determined and benchmark of sine gray-value distribution stripe pattern can be coded. First, using spin filtering can filter off the noise on stripe images, then using derivative-sign binary image method extract the stripe skeletons, and the bright and dark stripe are endowed with different color. So we can code all stripe with color alternate rule.
On the basis of analyzed tricolor-coded stripe pattern and sine gray-value distribution stripe pattern, tricolor-coded and sine gray-value distribution stripe pattern and the methods of stripe image procession are proposed.
On the end of the paper, the experimental results show that the new spin filtering provides optimal results in filtering off noise without distortion of the fringe features, regardless of fringe curvature. Experimental comparison of the methods of stripe image procession and stripe order recognition in the three stripe patterns is given. Finally, respective advantage and disadvantage and application range for the three stripe patterns are discussed.
本文研究了三基色条纹模板下条纹图像处理及条纹编码方法。三基色条纹模板利用颜色信息提高了光学三维测量的抗干扰能力,有利于条纹断裂处中心线对应条纹的寻找。提出了在该模板下的条纹图像处理算法中无需任何偏色图像先验信息或理想假设条件的图像偏色调整方法;改进了条件细化算法;研究了基于搜索线的条纹自动编码算法。实验表明得到的条纹中心线单像素程度高、居中性好、无扭曲变形、断点少;条纹自动编码结果正确,无漏编、错编现象,处理结果达到工程实际应用的要求。
本文研究了灰度正弦变化条纹模板下条纹图像处理及条纹编码方法。在灰度正弦变化条纹模板中以彩色中央条纹为标志条纹,确立条纹编码基准。利用旋滤波对条纹图像进行预处理,运用二维导数符号二值图法提取条纹中心线,并将所提取的亮条纹中心线与暗条纹中心线赋予不同颜色,以便后续处理中利用色彩的相间规律对条纹进行编码。
本文在分析三基色条纹模板和灰度正弦变化条纹模板各自特点的基础上,研究了三基色灰度正弦变化条纹投射模式及其条纹图像处理方法。
最后,对三种条纹模板下条纹图像处理及条纹编码的实验结果进行了对比分析,总结了三种条纹模板各自的优缺点和适用范围。
Non-contact 3-Dimensiongal optical model reconstruction is an important branch of modern optics, laser technique, digital image processing technique and machine vision, which is widely applied in the fields of reverse engineering, CAD/CAM, mold design, medical treatment, measurement online, individuation manufacturing and so on. The method called face-structure grating projection became a study hot point of 3-Dimensiongal optical model reconstruction. This method can measure the surface of 3D object rapidly and nicely. Further more, it has some advantages, such as non-contract measured process, high measure speed, high resolving ability of data space, little cost and so on. One of its key techniques is the digital stripe image processing. Based on comparing and analyzing the existing techniques, this paper deeply studied the stripe pattern of projecting grating.
The methods of stripe image procession and stripe order recognition in tricolor-coded stripe pattern are studied and proposed, and theoretical analysis and experimental results are given. Tricolor-coded stripe pattern has the advantages of providing information for stripe order recognition, and thereby carries the measurement system high ability of disturbance-resistance. An algorithm is proposed to automatically correct the color of any image without using any additional knowledge or any assumed ideal condition. This paper presents a stripe thinning algorithm by improving Condition Parallel Algorithm. An automatic stripe order recognition and storage method for tricolor-coded stripe projection pattern is also presented in the paper. Finally, the results of the experiments are also presented to prove the effectiveness of the above discussed methods.
The methods of stripe image procession and stripe order recognition in sine gray-value distribution stripe pattern are studied and proposed. Furthermore, theoretical analysis and experimental results are given. With the stripe color ,to determine the symbol stripe can be determined and benchmark of sine gray-value distribution stripe pattern can be coded. First, using spin filtering can filter off the noise on stripe images, then using derivative-sign binary image method extract the stripe skeletons, and the bright and dark stripe are endowed with different color. So we can code all stripe with color alternate rule.
On the basis of analyzed tricolor-coded stripe pattern and sine gray-value distribution stripe pattern, tricolor-coded and sine gray-value distribution stripe pattern and the methods of stripe image procession are proposed.
On the end of the paper, the experimental results show that the new spin filtering provides optimal results in filtering off noise without distortion of the fringe features, regardless of fringe curvature. Experimental comparison of the methods of stripe image procession and stripe order recognition in the three stripe patterns is given. Finally, respective advantage and disadvantage and application range for the three stripe patterns are discussed.
引文
[1]黄炎.基于立体视觉的脚型测量系统中的关键技术研究[D].杭州:浙江大学, 2006.
[2]张国雄.三坐标测量机[M].天津:天津大学出版社, 1999.
[3]张可.基于双目立体视觉原理的自由曲面三维重构[D].武汉:华中科技大学, 2005.
[4] Kanade T, Okutomi M. A Stereo Matching Algorithm with an Adaptive Window: Theory and Experiment[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence. 1994, 16(9): 920-932.
[5] Tsai P S, Shah M. Shape from shading using linear approximation[J]. Image and Vision Computing. 1994, 12(8): 487-498.
[6]胡岩峰,刘波,李峰,等.由光影明暗度获得物体表面形状的快速算法[J].光子学报. 2003, 32(008): 985-988.
[7] Brown C W. Dimensional inspection techniques for sample-point measurement technology[J]. Precision engineering. 1992, 14(2): 110-111.
[8]吴剑波,崔振.光刀中心自适应阈值提取法[J].半导体光电. 2001, 22(001): 62-64.
[9] Hosni Y, Ferreira L. Laser based system for reverse engineering[J]. Computers and Industrial Engineering. 1994, 26(2): 387-394.
[10]王朝阳,戴福隆.条纹图象的数字化自动分析处理技术之二:相位分析法[J].光子学报. 1999, 28(011): 996-1001.
[11]杨福俊,何小元,王蕴珊.一种新的单幅条纹图的相位解调方法[J].光子学报. 2005, 34(8): 1257-1261.
[12]尹志武,程维明,陈明仪.条纹图的图像处理方法[J].光学精密工程. 1999, 7(01): 51-58.
[13]戴福隆,王朝阳.条纹图象的数字化自动分析处理技术之一:条纹中心法[J].光子学报. 1999(08): 700-706.
[14]郝煜栋,赵洋.光学投影式三维轮廓测量技术综述[J].光学技术. 1998(005): 57-60.
[15] http://www.homometrica.ch[Z].
[16] http://www.cyberware.com[Z].
[17] http://www.vorum.com[Z].
[18]沈威,李德华,李和平,等.三维激光彩色扫描仪中展开图的获取及拼接方法研究[J].计算机工程与应用. 2005, 41(13): 29-31.
[19] http://www.breuckmann.com[Z].
[20]张吉华,庄葆华.激光三角法轮廓测量[J].天津大学学报. 1995, 28(2): 265-269.
[21]王晓林,陈伟民.光切法三维轮廓测量的原理及其应用[J].光学技术. 1997(002): 39-43.
[22]牛小兵,林玉池,赵美蓉,等.光栅投影三维测量的原理及关键技术分析[J].光电子激光. 2002, 13(9): 983-986.
[23]于起峰,陆宏伟,刘肖琳.基于图像的精密测量与运动测量[M].北京:科学出版社, 2002.
[24] Sansoni G, Corini S, Lazzari S, et al. Three-dimensional imaging based on Gray-code light projection: characterization of the measuring algorithm and development of a measuring system for industrial applications[J]. Appl. Opt. 1997, 36(23): 4463-4472.
[25]盖绍彦.结构光法形体检测系统中的条纹图像处理[D].南京:东南大学, 2004.
[26] Liu W, Wang Z, Mu G, et al. A novel profilometry with color-coded project grating and itsapplication in 3D reconstruction[J]. Communications, 1999. APCC/OECC'99. Fifth Asia-Pacific Conference on... and Fourth Optoelectronics and Communications Conference. 1999, 2.
[27]刘维一,王肇圻.彩色数字编码投影光栅三维轮廓术的研究[J].光学学报. 2001, 21(006): 687-690.
[28] Hausler G, Ritter D. Parallel three-dimensional sensing by color-coded triangulation[J]. Applied Optics. 1993, 32(35): 7164-7169.
[29]吕乃光,冯迪,邓文怡,等.用BP神经网络构成三维视觉测量系统的新方法[J].华中科技大学学报(自然科学版). 2002, 30(5): 19-21.
[30]郑建铧,郝重阳,雷方元,等.利用色彩直方图特征进行偏色图象的自动检测和校正[J].中国图象图形学报. 2003, 08(09).
[31] Gasparini F, Schettini R. Color correction for digital photographs[J]. Image Analysis and Processing, 2003. Proceedings. 12th International Conference on. 2003: 646-651.
[32]谷元保,付宇卓.一种基于灰度世界模型自动白平衡方法[J].计算机仿真. 2005, 22(9): 185-188.
[33] Otsu N. A threshold selection method from gray level[J]. IEEE Transactions on Systems, Man, and Cybernetics. 1979, 9(1): 62-66.
[34]何斌,马天予,王运坚. Visual C++数字图像处理[M].北京:人民邮电出版社, 2001.
[35]刘志敏,杨杰.基于数学形态学的图像形态滤波[J].红外与激光工程. 1999, 28(04): 10-15.
[36] Hilditch C J. Comparison of thinning algorithms on a parallel processor[J]. Image Vision Comput. 1983, 1(3): 115-132.
[37] Deutsch E S. Thinning algorithms on rectangular, hexagonal, and triangular arrays[J].Communications of the ACM. 1972, 15(9): 827-837.
[38] Zhang T Y, Suen C Y. A fast parallel algorithm for thinning digital patterns[J]. Communications of the ACM. 1984, 27(3): 236-239.
[39] Yu Q, Liu X, Andresen K. New spin filters for interferometric fringe patterns and grating patterns[J]. Applied Optics. 1994, 33(17): 3705-3711.
[40] Yu Q, Andresen K. Fringe orientation maps and 2D derivative-sign binary image methods for extraction of fringe skeletons[J]. Applied. Optics. 1994, 33(29): 6873–6878.
[41] F Y Q. Spin Filtering Processes and Automatical Extraction of Fringe Centerlines in Digital Image of Interferometric Fringes[J]. Applied Optics. 1988, 27(18): 3782-3784.
[2]张国雄.三坐标测量机[M].天津:天津大学出版社, 1999.
[3]张可.基于双目立体视觉原理的自由曲面三维重构[D].武汉:华中科技大学, 2005.
[4] Kanade T, Okutomi M. A Stereo Matching Algorithm with an Adaptive Window: Theory and Experiment[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence. 1994, 16(9): 920-932.
[5] Tsai P S, Shah M. Shape from shading using linear approximation[J]. Image and Vision Computing. 1994, 12(8): 487-498.
[6]胡岩峰,刘波,李峰,等.由光影明暗度获得物体表面形状的快速算法[J].光子学报. 2003, 32(008): 985-988.
[7] Brown C W. Dimensional inspection techniques for sample-point measurement technology[J]. Precision engineering. 1992, 14(2): 110-111.
[8]吴剑波,崔振.光刀中心自适应阈值提取法[J].半导体光电. 2001, 22(001): 62-64.
[9] Hosni Y, Ferreira L. Laser based system for reverse engineering[J]. Computers and Industrial Engineering. 1994, 26(2): 387-394.
[10]王朝阳,戴福隆.条纹图象的数字化自动分析处理技术之二:相位分析法[J].光子学报. 1999, 28(011): 996-1001.
[11]杨福俊,何小元,王蕴珊.一种新的单幅条纹图的相位解调方法[J].光子学报. 2005, 34(8): 1257-1261.
[12]尹志武,程维明,陈明仪.条纹图的图像处理方法[J].光学精密工程. 1999, 7(01): 51-58.
[13]戴福隆,王朝阳.条纹图象的数字化自动分析处理技术之一:条纹中心法[J].光子学报. 1999(08): 700-706.
[14]郝煜栋,赵洋.光学投影式三维轮廓测量技术综述[J].光学技术. 1998(005): 57-60.
[15] http://www.homometrica.ch[Z].
[16] http://www.cyberware.com[Z].
[17] http://www.vorum.com[Z].
[18]沈威,李德华,李和平,等.三维激光彩色扫描仪中展开图的获取及拼接方法研究[J].计算机工程与应用. 2005, 41(13): 29-31.
[19] http://www.breuckmann.com[Z].
[20]张吉华,庄葆华.激光三角法轮廓测量[J].天津大学学报. 1995, 28(2): 265-269.
[21]王晓林,陈伟民.光切法三维轮廓测量的原理及其应用[J].光学技术. 1997(002): 39-43.
[22]牛小兵,林玉池,赵美蓉,等.光栅投影三维测量的原理及关键技术分析[J].光电子激光. 2002, 13(9): 983-986.
[23]于起峰,陆宏伟,刘肖琳.基于图像的精密测量与运动测量[M].北京:科学出版社, 2002.
[24] Sansoni G, Corini S, Lazzari S, et al. Three-dimensional imaging based on Gray-code light projection: characterization of the measuring algorithm and development of a measuring system for industrial applications[J]. Appl. Opt. 1997, 36(23): 4463-4472.
[25]盖绍彦.结构光法形体检测系统中的条纹图像处理[D].南京:东南大学, 2004.
[26] Liu W, Wang Z, Mu G, et al. A novel profilometry with color-coded project grating and itsapplication in 3D reconstruction[J]. Communications, 1999. APCC/OECC'99. Fifth Asia-Pacific Conference on... and Fourth Optoelectronics and Communications Conference. 1999, 2.
[27]刘维一,王肇圻.彩色数字编码投影光栅三维轮廓术的研究[J].光学学报. 2001, 21(006): 687-690.
[28] Hausler G, Ritter D. Parallel three-dimensional sensing by color-coded triangulation[J]. Applied Optics. 1993, 32(35): 7164-7169.
[29]吕乃光,冯迪,邓文怡,等.用BP神经网络构成三维视觉测量系统的新方法[J].华中科技大学学报(自然科学版). 2002, 30(5): 19-21.
[30]郑建铧,郝重阳,雷方元,等.利用色彩直方图特征进行偏色图象的自动检测和校正[J].中国图象图形学报. 2003, 08(09).
[31] Gasparini F, Schettini R. Color correction for digital photographs[J]. Image Analysis and Processing, 2003. Proceedings. 12th International Conference on. 2003: 646-651.
[32]谷元保,付宇卓.一种基于灰度世界模型自动白平衡方法[J].计算机仿真. 2005, 22(9): 185-188.
[33] Otsu N. A threshold selection method from gray level[J]. IEEE Transactions on Systems, Man, and Cybernetics. 1979, 9(1): 62-66.
[34]何斌,马天予,王运坚. Visual C++数字图像处理[M].北京:人民邮电出版社, 2001.
[35]刘志敏,杨杰.基于数学形态学的图像形态滤波[J].红外与激光工程. 1999, 28(04): 10-15.
[36] Hilditch C J. Comparison of thinning algorithms on a parallel processor[J]. Image Vision Comput. 1983, 1(3): 115-132.
[37] Deutsch E S. Thinning algorithms on rectangular, hexagonal, and triangular arrays[J].Communications of the ACM. 1972, 15(9): 827-837.
[38] Zhang T Y, Suen C Y. A fast parallel algorithm for thinning digital patterns[J]. Communications of the ACM. 1984, 27(3): 236-239.
[39] Yu Q, Liu X, Andresen K. New spin filters for interferometric fringe patterns and grating patterns[J]. Applied Optics. 1994, 33(17): 3705-3711.
[40] Yu Q, Andresen K. Fringe orientation maps and 2D derivative-sign binary image methods for extraction of fringe skeletons[J]. Applied. Optics. 1994, 33(29): 6873–6878.
[41] F Y Q. Spin Filtering Processes and Automatical Extraction of Fringe Centerlines in Digital Image of Interferometric Fringes[J]. Applied Optics. 1988, 27(18): 3782-3784.