多尺度自适应边缘检测方法研究
|
摘要
边缘检测技术是图像测量、图像分割、图像压缩以及模式识别等图像处理技术的基础，是计算机视觉和图像测量重要的研究课题之一。如何快速、准确的对图像进行边缘检测是图像处理领域的研究热点。
本文首先简要介绍了经典边缘检测算子和一些成熟的边缘检测方法，论述了这些算法的基本原理，比较和分析了它们的优缺点。
本文重点论述了图像处理的多尺度自适应滤波技术。介绍了一种根据局部灰度变化信息，估计边缘尺度的多尺度自适应高斯滤波方法。本文详细论述了多尺度自适应高斯滤波的基本理论和实现方法，并对这种方法进行了有效的改进。
本文提出了一种基于Canny理论，利用多尺度自适应高斯滤波技术的Canny自适应边缘检测方法。Canny自适应边缘检测方法在保持了Canny算子原有的定位准确，单边缘响应和信噪比高等优点的基础上，提高了Canny算子在提取图像边缘细节信息和抑制假边缘方面的性能。Canny自适应边缘检测方法利用图像梯度特征信息自适应的生成动态阈值，提高了边缘检测的自动化程度。本文详细介绍了这种算法的基本思想和具体方法，并通过实验以及对实验结果的分析，证明了Canny自适应边缘检测方法是一种有效的边缘检测方法。
结合论文所述的多尺度自适应高斯滤波方法和Canny自适应边缘检测方法，在实验中实现了对测量图像中标定板的定位、标定点的边缘轮廓和中心坐标的精确定位。通过实验数据和结果，证明了使用本文提出的Canny自适应边缘检测方法可以从不同尺度上快速、准确的检测出目标的边缘轮廓。
Edge detection is the base of the image measuring, image segmentation, image compressing and pattern recognition and so on. Edge detection is one of the important subjects in the field of computer vision and image measuring. How to detect the edge of object exactly and quickly is a researchful holspot of image processing.
Some classic methods and efficient edge-detection methods are introduced in this paper. The basic concepts of these techniques are summarized and the methods' advantages and disadvantages are analyzed.
Multiscale adaptive filtering is particularly discussed in this paper. Multiscale adaptive Gauss filtering is proposed. This method estimates the scale of edge based on local gray scale statistic. The basic theory and the concrete algorithm of the multiscale adaptive Gauss filtering are summarized and some improvements are made.
Based on Canny theory and multiscale adaptive Gauss filtering, an adaptive Canny edge-detection method is proposed in this paper. This method not only keeps the cxcellent performance of Canny operator in good localization, only one response to a single edge and good detection, but also improves the performance in the detail edge-detection and good detection. Canny adaptive method detects the edges with adaptive threshold value according to the gradient information. That improves the automaticity of edge-detection. The theories and the concrete algorithm of the Canny adaptive method are introduced in this paper. With the mathematic analysis and test result, it demonstrates that the Canny adaptive edge-detection method is a valid method on edge-detection.
Using the multiscale adaptive Gauss filtering method and the Canny adaptive method, a experiment is designed to orientate the surveyor's board, extract the exact edges and the center coordinate figures of the surveyor points. With the mathematic analysis and lots of experiments, it demonstrates that using the method proposed in this paper can quickly and exactly detect the edge of object in different scales.
本文首先简要介绍了经典边缘检测算子和一些成熟的边缘检测方法，论述了这些算法的基本原理，比较和分析了它们的优缺点。
本文重点论述了图像处理的多尺度自适应滤波技术。介绍了一种根据局部灰度变化信息，估计边缘尺度的多尺度自适应高斯滤波方法。本文详细论述了多尺度自适应高斯滤波的基本理论和实现方法，并对这种方法进行了有效的改进。
本文提出了一种基于Canny理论，利用多尺度自适应高斯滤波技术的Canny自适应边缘检测方法。Canny自适应边缘检测方法在保持了Canny算子原有的定位准确，单边缘响应和信噪比高等优点的基础上，提高了Canny算子在提取图像边缘细节信息和抑制假边缘方面的性能。Canny自适应边缘检测方法利用图像梯度特征信息自适应的生成动态阈值，提高了边缘检测的自动化程度。本文详细介绍了这种算法的基本思想和具体方法，并通过实验以及对实验结果的分析，证明了Canny自适应边缘检测方法是一种有效的边缘检测方法。
结合论文所述的多尺度自适应高斯滤波方法和Canny自适应边缘检测方法，在实验中实现了对测量图像中标定板的定位、标定点的边缘轮廓和中心坐标的精确定位。通过实验数据和结果，证明了使用本文提出的Canny自适应边缘检测方法可以从不同尺度上快速、准确的检测出目标的边缘轮廓。
Edge detection is the base of the image measuring, image segmentation, image compressing and pattern recognition and so on. Edge detection is one of the important subjects in the field of computer vision and image measuring. How to detect the edge of object exactly and quickly is a researchful holspot of image processing.
Some classic methods and efficient edge-detection methods are introduced in this paper. The basic concepts of these techniques are summarized and the methods' advantages and disadvantages are analyzed.
Multiscale adaptive filtering is particularly discussed in this paper. Multiscale adaptive Gauss filtering is proposed. This method estimates the scale of edge based on local gray scale statistic. The basic theory and the concrete algorithm of the multiscale adaptive Gauss filtering are summarized and some improvements are made.
Based on Canny theory and multiscale adaptive Gauss filtering, an adaptive Canny edge-detection method is proposed in this paper. This method not only keeps the cxcellent performance of Canny operator in good localization, only one response to a single edge and good detection, but also improves the performance in the detail edge-detection and good detection. Canny adaptive method detects the edges with adaptive threshold value according to the gradient information. That improves the automaticity of edge-detection. The theories and the concrete algorithm of the Canny adaptive method are introduced in this paper. With the mathematic analysis and test result, it demonstrates that the Canny adaptive edge-detection method is a valid method on edge-detection.
Using the multiscale adaptive Gauss filtering method and the Canny adaptive method, a experiment is designed to orientate the surveyor's board, extract the exact edges and the center coordinate figures of the surveyor points. With the mathematic analysis and lots of experiments, it demonstrates that using the method proposed in this paper can quickly and exactly detect the edge of object in different scales.
引文
[1]Rafael C. Gonzalez and Richard E. Woods. 数字图像处理(第2版)．北京：电子工业出版社．2003．463-467
[2]L. G. Roberts. Machine perception of three-dimensional solids. In J T Tippett, editor, Optical and Electro—Optical Information Processing, New York:MIT Press, 1982, 159-197
[3]D. Marr, E. Hildreth. Theory of edge detection. Proceedings of the Royal Society, 1980, 187-217
[4]D. Marr, E. Hildreth. Theory of edge detection. Computer Vision, 1991, 77-107
[5]俞妍妍，王继成．改进的多尺度形态梯度的图像边缘检测．计算机工程与应用，2003(18)：75-82
[6]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．光电子·激光，2003，14(8)：862-865
[7]周焰，王祖喜，李德华．基于弧长参数的曲线自适应平滑算法．红外与激光工程，2000(1)：25-29
[8]尹平，王润生．自适应多尺度边缘检测．软件学报，2000，11(8)：990-994
[9]李翠华，郑南宁，张永平．基于低松弛迭代格式的快速自适应滤波算法．西安交通大学学报，1998，32(11)：1-4
[10]陈少平，朱翠涛．基于振幅谱重构信号的松弛迭代算法．中南民族大学(自然科学版)，2002(4)：30-33
[11]J. Canny. A Computational Approach to Edge Detection. IEEETransactions on Patten Analysis and Machine Intelligence, 1986,8(6):679-698
[12]A. Rosenfeld, M. Thurston. Edge and curve detection for visual scene analysis. IEEE Transactions on Computers, 1971,20(5):562-569
[13]C.J. Vianna, E.A. De Souza, Combining Mart and Canny operators to detect edges. Mirowave and Optoelectronics Conference, 2003,2(9):695-699
[14]Hongjian Shi, R. Ward. Canny edge based image expansion. Circuits and Systems, 2002(1):785-788
[15]Kar-Kin Lee, Wai-Kuen Cham, Qin-Ran Chen. Chin contour estimation using modified Canny edge detector. Control, Automation, Robotics and Vision, 2002 (2) : 770-775
[16]Mohamed All, David Clausi. Using the Canny Edge Detector for Feature Extraction and Enhancement of Remote Sensing Images. Geoscience and Remote Sensing Symposium, 2001, 5(7) :9-13
[17]F. Bergho]m. Edge Focusing. IEEE Trans., 1987, 9(6) :726-741
[18]G. Heene, S. Gautama. Optimisation of a coastline extraction algorithm, for object-oriented matching of multisensor satellite imagery. Geoscience and Remote Sensing Symposium, 2000, Proceedings, GARSS 2000, IEEE International, 2000, 6 (7): 26322-2634
[19]赵西安，李德仁．小波多尺度影像边缘聚焦算法．遥感学报，2004(4)：301-303
[20]F. Faghih, M. Smith. Combining spatial and scale-space techniques for edge detection to provide a spatially adaptive wavelet-based noise filtering algorithm. Image Processing, IEEE Transactions on, 2002,11(9):1062-1071
[21]陆军，王润生．一种基于尺度空间理论的直线抽取算法．中国图像图形学报，2002(8)：693-696
[22]韩军伟．景物边缘提取技术的研究与应用[学位论文]．西安：西北工业大学，2001．19-22
[23]J H Elder ,S W Zucker. Local scale control for edge detection and blur estimation. In B Buxton and R Cipolla, editors, 4th European Conference on Computer Vision, Cambridge, England, 1996, 57-69
[24]T Aydin, Y Yemez, E Anarim, B R Hunt. Multidirectional and multiscale edge detection via M—band wavelet transform. IEEE Transactions on Image Processing, 1996(5):1370-1377
[25]T Lindeberg. Edge detection and ridge detection with automatic scale selection. 1966 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1996,465-470
[26]M. Acharyya, M. K. Kundu, Document image segmentation using wavelet scale-space features. IEEE Transactions on Circuits and Systems for Video Technology, 2002,12(10):1117-1127
[27]F. Faghih, M. Smith, Combining spatial and scale-space techniques for edge detection to provide a spatially adaptive wavelet-based noise filtering algorithm. IEEE Transactions on Image Processing, 2002,11(9):1062-1071
[28]Ming Tang, Songde Ma. Corrections to general scheme of region competition based on scale space. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002,24(?):1007-1007
[29]贾云得．机器视觉．北京：科学出版社，2000．71-78
[30]A.L. Yuille ,T. Poggio, Fingerprint theorems for zero crossing. Opt. Soc. Am. A, 1985(2) :683-692
[31]A.L. Yuille ,T. Poggio, Scaling theorems for zero crossing. IEEE Transaction on Pattern Analysis and Machine Intelligence, 1986,8(1):15-25
[32]Hong Jeong, C. I. Kim. Adaptive Determination of Filter Scales for Edge Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14 (5) :579-585
[33]Deok J. Park, Kwon M. Nam and Rae—Hong Park. Multiresolution Edge Detection Techniques. Pattern Recognition. 1994, 28(2):211-229
[34]Mallat. D. A Compact Multiresolution Representation: The Wavelet Model. 1987, Proc. IEEE Computer Society Workshop on Computer Vision, Washington D.C.:IEEE Computer Society Press, 198?,2-7
[35]尹平，王润生．多分辨率图像锥的产生方法及分析．计算机工程与科学．1999，21(6)：35-39
[36]D. Demigny and T. Kamle. A discrete expression of Canny's criteria for step edge detector performances evaluation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1997,19(11):1199-1211
[37]D. Demigny. Extension of Canny's discrete criteria to second derivative filters, towards a unified approach. 1998 International Conference on Image Processing, 1998(2):520-524
[38]Worthington. P. L. Enhanced. Canny edge detection using curvature consistency. 2002 Proceedings. 16th International Conference on Pattern Recognition, 2002,1(11):596-599
[39]Gang Liu, R.M. Haralick, Two practical issues in Canny's edge detector implementation. Pattern Recognition, 2000(3):676-678
[40]杨枝灵，王开等．Visual C++数字图像获取处理及实践应用．北京：人民邮电出版社，2003．568-569
[41]Louisa Lam, Seong-Whan Lee. Thinning Methodologies-A Comprehensive Survey, IEEE Transactions of Pattern analysis and Machine Intelligence, 1992,14(9):869-885
[42]万力，易昂，傅明．一种基于Canny算法的边缘提取改善方法．计算机技术与自动化，2003，22(1)：24-26
[43]于起峰，陆宏伟，刘肖琳．基于图像的精密测量与运动测量．北京：科学出版社，2002．135-139
[2]L. G. Roberts. Machine perception of three-dimensional solids. In J T Tippett, editor, Optical and Electro—Optical Information Processing, New York:MIT Press, 1982, 159-197
[3]D. Marr, E. Hildreth. Theory of edge detection. Proceedings of the Royal Society, 1980, 187-217
[4]D. Marr, E. Hildreth. Theory of edge detection. Computer Vision, 1991, 77-107
[5]俞妍妍，王继成．改进的多尺度形态梯度的图像边缘检测．计算机工程与应用，2003(18)：75-82
[6]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．光电子·激光，2003，14(8)：862-865
[7]周焰，王祖喜，李德华．基于弧长参数的曲线自适应平滑算法．红外与激光工程，2000(1)：25-29
[8]尹平，王润生．自适应多尺度边缘检测．软件学报，2000，11(8)：990-994
[9]李翠华，郑南宁，张永平．基于低松弛迭代格式的快速自适应滤波算法．西安交通大学学报，1998，32(11)：1-4
[10]陈少平，朱翠涛．基于振幅谱重构信号的松弛迭代算法．中南民族大学(自然科学版)，2002(4)：30-33
[11]J. Canny. A Computational Approach to Edge Detection. IEEETransactions on Patten Analysis and Machine Intelligence, 1986,8(6):679-698
[12]A. Rosenfeld, M. Thurston. Edge and curve detection for visual scene analysis. IEEE Transactions on Computers, 1971,20(5):562-569
[13]C.J. Vianna, E.A. De Souza, Combining Mart and Canny operators to detect edges. Mirowave and Optoelectronics Conference, 2003,2(9):695-699
[14]Hongjian Shi, R. Ward. Canny edge based image expansion. Circuits and Systems, 2002(1):785-788
[15]Kar-Kin Lee, Wai-Kuen Cham, Qin-Ran Chen. Chin contour estimation using modified Canny edge detector. Control, Automation, Robotics and Vision, 2002 (2) : 770-775
[16]Mohamed All, David Clausi. Using the Canny Edge Detector for Feature Extraction and Enhancement of Remote Sensing Images. Geoscience and Remote Sensing Symposium, 2001, 5(7) :9-13
[17]F. Bergho]m. Edge Focusing. IEEE Trans., 1987, 9(6) :726-741
[18]G. Heene, S. Gautama. Optimisation of a coastline extraction algorithm, for object-oriented matching of multisensor satellite imagery. Geoscience and Remote Sensing Symposium, 2000, Proceedings, GARSS 2000, IEEE International, 2000, 6 (7): 26322-2634
[19]赵西安，李德仁．小波多尺度影像边缘聚焦算法．遥感学报，2004(4)：301-303
[20]F. Faghih, M. Smith. Combining spatial and scale-space techniques for edge detection to provide a spatially adaptive wavelet-based noise filtering algorithm. Image Processing, IEEE Transactions on, 2002,11(9):1062-1071
[21]陆军，王润生．一种基于尺度空间理论的直线抽取算法．中国图像图形学报，2002(8)：693-696
[22]韩军伟．景物边缘提取技术的研究与应用[学位论文]．西安：西北工业大学，2001．19-22
[23]J H Elder ,S W Zucker. Local scale control for edge detection and blur estimation. In B Buxton and R Cipolla, editors, 4th European Conference on Computer Vision, Cambridge, England, 1996, 57-69
[24]T Aydin, Y Yemez, E Anarim, B R Hunt. Multidirectional and multiscale edge detection via M—band wavelet transform. IEEE Transactions on Image Processing, 1996(5):1370-1377
[25]T Lindeberg. Edge detection and ridge detection with automatic scale selection. 1966 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1996,465-470
[26]M. Acharyya, M. K. Kundu, Document image segmentation using wavelet scale-space features. IEEE Transactions on Circuits and Systems for Video Technology, 2002,12(10):1117-1127
[27]F. Faghih, M. Smith, Combining spatial and scale-space techniques for edge detection to provide a spatially adaptive wavelet-based noise filtering algorithm. IEEE Transactions on Image Processing, 2002,11(9):1062-1071
[28]Ming Tang, Songde Ma. Corrections to general scheme of region competition based on scale space. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002,24(?):1007-1007
[29]贾云得．机器视觉．北京：科学出版社，2000．71-78
[30]A.L. Yuille ,T. Poggio, Fingerprint theorems for zero crossing. Opt. Soc. Am. A, 1985(2) :683-692
[31]A.L. Yuille ,T. Poggio, Scaling theorems for zero crossing. IEEE Transaction on Pattern Analysis and Machine Intelligence, 1986,8(1):15-25
[32]Hong Jeong, C. I. Kim. Adaptive Determination of Filter Scales for Edge Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14 (5) :579-585
[33]Deok J. Park, Kwon M. Nam and Rae—Hong Park. Multiresolution Edge Detection Techniques. Pattern Recognition. 1994, 28(2):211-229
[34]Mallat. D. A Compact Multiresolution Representation: The Wavelet Model. 1987, Proc. IEEE Computer Society Workshop on Computer Vision, Washington D.C.:IEEE Computer Society Press, 198?,2-7
[35]尹平，王润生．多分辨率图像锥的产生方法及分析．计算机工程与科学．1999，21(6)：35-39
[36]D. Demigny and T. Kamle. A discrete expression of Canny's criteria for step edge detector performances evaluation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1997,19(11):1199-1211
[37]D. Demigny. Extension of Canny's discrete criteria to second derivative filters, towards a unified approach. 1998 International Conference on Image Processing, 1998(2):520-524
[38]Worthington. P. L. Enhanced. Canny edge detection using curvature consistency. 2002 Proceedings. 16th International Conference on Pattern Recognition, 2002,1(11):596-599
[39]Gang Liu, R.M. Haralick, Two practical issues in Canny's edge detector implementation. Pattern Recognition, 2000(3):676-678
[40]杨枝灵，王开等．Visual C++数字图像获取处理及实践应用．北京：人民邮电出版社，2003．568-569
[41]Louisa Lam, Seong-Whan Lee. Thinning Methodologies-A Comprehensive Survey, IEEE Transactions of Pattern analysis and Machine Intelligence, 1992,14(9):869-885
[42]万力，易昂，傅明．一种基于Canny算法的边缘提取改善方法．计算机技术与自动化，2003，22(1)：24-26
[43]于起峰，陆宏伟，刘肖琳．基于图像的精密测量与运动测量．北京：科学出版社，2002．135-139