自动图像配准算法研究
|
摘要
自动图像配准是以自动化的方式实现对图像的高精度、高效率配准的综合性图像处理技术。作为图像处理的核心技术,自动图像配准是图像融合、图像重建、三维建模等的关键性技术之一,在工业、军事、科研等领域有着广泛的应用前景和极高的应用价值。
本文详细探讨了图像配准中几何变换模型、特征检测匹配算法以及模型参数估计算法。为了满足高精高效的要求,本文对SIFT算法和RANSAC算法都做了大量修改。但自动图像配准不是这些算法的简单组合应用,它有着完整的体系结构。从系统的角度出发,本文引入了图像多尺度分析技术和相关区域分析技术,提出了“自下而上,逐层迭代,逐步求精”的思想。图像多尺度分析消除了图像中的细节信息,使得检测匹配的特征点更加稳定。相关区域分析检测待配准图像间的重叠区域,逐层迭代时只检测相关区域,使得自动配准更高效。综合运用这些图像处理技术,形成了完整有机的自动图像配准的体系结构。
仿真实验结果表明,本文提出的自动图像配准算法的估算参数精度高,误差小,基本不受待配准图像尺寸影响。分组实验表明,自动图像配准算法的运行时间较直接配准方法有了极大的改进,实现了算法的高效率。故本文研究的自动图像配准算法基本实现了自动化、高精度、高效率的三大要求。
Automatic image registration is a comprehensive image processing technology that to achieve the high precision and high efficiency image registration automatically. As the core technology of image processing, automatic image registration is one of the key methods in image fusion, image reconstruction, 3D modeling. And it has wide application and high value n the industrial, military, scientific research.
This paper discusses the geometric transformation model, feature detection and matching algorithms, and model parameter estimation algorithm in detail. In order to meet the requirements of high precision and efficient, SIFT algorithm and RANSAC algorithm have been changed a lot. But automatic image registration is not the simple combination of these algorithms; it has a complete architecture of itself. From the view of system, this paper introduces the image multiscale analysis technology and related areas analysis technology, proposed method of "bottom to up, layer by layer, step by step". Image multiscale analysis eliminates the image details, and let image feature points be more stable. Related areas analysis detects overlap areas between images, and makes automatic registration be more efficient. Automatic image registration forms a complete organic architecture by combination of these image processing techniques.
Simulation results show that the proposed automatic image registration algorithm in this paper can estimate parameters in high precision, small error, and can't affect by the image size. Compared by direct registration method, Group results show that automatic image registration algorithm has been greatly improved, achieved high efficiency. Therefore, this paper's automatic image registration algorithm achieves three requirements of automated, high precision and high efficiency.
本文详细探讨了图像配准中几何变换模型、特征检测匹配算法以及模型参数估计算法。为了满足高精高效的要求,本文对SIFT算法和RANSAC算法都做了大量修改。但自动图像配准不是这些算法的简单组合应用,它有着完整的体系结构。从系统的角度出发,本文引入了图像多尺度分析技术和相关区域分析技术,提出了“自下而上,逐层迭代,逐步求精”的思想。图像多尺度分析消除了图像中的细节信息,使得检测匹配的特征点更加稳定。相关区域分析检测待配准图像间的重叠区域,逐层迭代时只检测相关区域,使得自动配准更高效。综合运用这些图像处理技术,形成了完整有机的自动图像配准的体系结构。
仿真实验结果表明,本文提出的自动图像配准算法的估算参数精度高,误差小,基本不受待配准图像尺寸影响。分组实验表明,自动图像配准算法的运行时间较直接配准方法有了极大的改进,实现了算法的高效率。故本文研究的自动图像配准算法基本实现了自动化、高精度、高效率的三大要求。
Automatic image registration is a comprehensive image processing technology that to achieve the high precision and high efficiency image registration automatically. As the core technology of image processing, automatic image registration is one of the key methods in image fusion, image reconstruction, 3D modeling. And it has wide application and high value n the industrial, military, scientific research.
This paper discusses the geometric transformation model, feature detection and matching algorithms, and model parameter estimation algorithm in detail. In order to meet the requirements of high precision and efficient, SIFT algorithm and RANSAC algorithm have been changed a lot. But automatic image registration is not the simple combination of these algorithms; it has a complete architecture of itself. From the view of system, this paper introduces the image multiscale analysis technology and related areas analysis technology, proposed method of "bottom to up, layer by layer, step by step". Image multiscale analysis eliminates the image details, and let image feature points be more stable. Related areas analysis detects overlap areas between images, and makes automatic registration be more efficient. Automatic image registration forms a complete organic architecture by combination of these image processing techniques.
Simulation results show that the proposed automatic image registration algorithm in this paper can estimate parameters in high precision, small error, and can't affect by the image size. Compared by direct registration method, Group results show that automatic image registration algorithm has been greatly improved, achieved high efficiency. Therefore, this paper's automatic image registration algorithm achieves three requirements of automated, high precision and high efficiency.
引文
[1] Beyer, W. H. (Ed.). CRC Standard Mathematical Tables, 28th ed. Boca Raton, FL: CRC Press, p. 211, 1987.
[2] Gray, A. Modern Differential Geometry of Curves and Surfaces with Mathematica, 2nd ed. Boca Raton, FL: CRC Press, p. 130, 1997.
[3] Zwillinger, D. (Ed.). "Affine Transformations."§4.3.2 in CRC Standard Mathematical Tables and Formulae. Boca Raton, FL: CRC Press, pp. 265-266, 1995.
[4] Kadison, L. and Kromann, M. T. Projective Geometry and Modern Algebra. Boston, MA: Birkh?user, 1996.
[5] Semple, J. G. Algebraic Projective Geometry. Oxford, England: Oxford University Press, 1998.
[6] David G. Lowe, "Object recognition from local scale-invariant features," International Conference on Computer Vision, Corfu, Greece (September 1999), pp. 1150-1157.
[7] David G. Lowe, "Distinctive image features from scale-invariant keypoints," International Journal of Computer Vision, 60, 2 (2004), pp. 91-110.
[8] Lindeberg, T. 1994. Scale-space theory: A basic tool for analysing structures at different scales. Journal of Applied Statistics, 21(2):224-270.
[9] Mikolajczyk, K. 2002. Detection of local features invariant to affine transformations, Ph.D. thesis, Institut National Polytechnique de Grenoble, France.
[10] Mikolajczyk, K., and Schmid, C. 2002. An affine invariant interest point detector. In European Conference on Computer Vision (ECCV), Copenhagen, Denmark, pp. 128-142.
[11] Brown, M. and Lowe, D.G. 2002. Invariant features from interest point groups. In British Machine Vision Conference, Cardiff, Wales, pp. 656-665.
[12] Harris, C. and Stephens, M. 1988. A combined corner and edge detector. In Fourth Alvey Vision Conference, Manchester, UK, pp. 147-151.
[13] Friedman, J.H., Bentley, J.L. and Finkel, R.A. 1977. An algorithm for finding best matches in logarithmic expected time. ACM Transactions on Mathematical Software, 3(3):209-226.
[14] Beis, J. and Lowe, D.G. 1997. Shape indexing using approximate nearest-neighbour search in high dimensional spaces. In Conference on Computer Vision and Pattern Recognition, Puerto Rico, pp. 1000-1006.
[15] Arya, S., and Mount, D.M. 1993. Approximate nearest neighbor queries in fixed dimensions. In Fourth Annual ACM-SIAM Symposium on Discrete Algorithms (SODA’93), pp. 271-280.
[16] Arya, S., Mount, D.M., Netanyahu, N.S., Silverman, R., and Wu, A.Y. 1998. An optimal algorithm for approximate nearest neighbor searching. Journal of the ACM, 45:891-923.
[17] M. A. Fischler and R. C. Bolles. Random sample consensus: A paradigm for model fittingith applications to image analysis and automated cartography. Communications of the ACM, 24:381–395, 1981.
[18] P. J. Huber. Robust Statistics. Wiley, 1981.
[19] P. J. Rousseeuw and A. M. Leroy. Robust Regression and Outlier Detection. Wiley, 1987.
[20] J. Mendel. Lessons in Estimation Theory for Signal Processing, Communication and Control. Prentice-Hall, Englewood-Cliffs, 1995.
[21] Z. Zhang. Parameter estimation techniques: A tutorial with application to conic fitting. Image and Vision Computing Journal, 25(1):59–76, 1997.
[22] C. V. Stewart. Robust parameter estimation in computer vision. SIAM Review, 41(3):513–537, September 1999.
[23] C. V. Stewart. MINPRAN: A new robust estimator for computer vision. IEEE Transactions on Pattern Analysis and Machine Intelligence, 17(10):925–938, October 1995.
[24] R.Hartley and A.Zisserman. Multiple View Geometry in Computer Vision. Cambridge University Press, second edition, 2003.
[25] B. J. Tordoff and D. W. Murray. Guided-MLESAC: Faster image transform estimation by using matching priors. IEEE Transactions on Pattern Analysis and Machine Intelligence, 27(10):1523–1535, October 2005.
[26] P. Burt and E. Adelson. A Multiresolution spline with Application to Image Mosaics. ACM Transactions on Graphics, Vol. 2, No. 4, October 1983. Pg. 217-236.
[2] Gray, A. Modern Differential Geometry of Curves and Surfaces with Mathematica, 2nd ed. Boca Raton, FL: CRC Press, p. 130, 1997.
[3] Zwillinger, D. (Ed.). "Affine Transformations."§4.3.2 in CRC Standard Mathematical Tables and Formulae. Boca Raton, FL: CRC Press, pp. 265-266, 1995.
[4] Kadison, L. and Kromann, M. T. Projective Geometry and Modern Algebra. Boston, MA: Birkh?user, 1996.
[5] Semple, J. G. Algebraic Projective Geometry. Oxford, England: Oxford University Press, 1998.
[6] David G. Lowe, "Object recognition from local scale-invariant features," International Conference on Computer Vision, Corfu, Greece (September 1999), pp. 1150-1157.
[7] David G. Lowe, "Distinctive image features from scale-invariant keypoints," International Journal of Computer Vision, 60, 2 (2004), pp. 91-110.
[8] Lindeberg, T. 1994. Scale-space theory: A basic tool for analysing structures at different scales. Journal of Applied Statistics, 21(2):224-270.
[9] Mikolajczyk, K. 2002. Detection of local features invariant to affine transformations, Ph.D. thesis, Institut National Polytechnique de Grenoble, France.
[10] Mikolajczyk, K., and Schmid, C. 2002. An affine invariant interest point detector. In European Conference on Computer Vision (ECCV), Copenhagen, Denmark, pp. 128-142.
[11] Brown, M. and Lowe, D.G. 2002. Invariant features from interest point groups. In British Machine Vision Conference, Cardiff, Wales, pp. 656-665.
[12] Harris, C. and Stephens, M. 1988. A combined corner and edge detector. In Fourth Alvey Vision Conference, Manchester, UK, pp. 147-151.
[13] Friedman, J.H., Bentley, J.L. and Finkel, R.A. 1977. An algorithm for finding best matches in logarithmic expected time. ACM Transactions on Mathematical Software, 3(3):209-226.
[14] Beis, J. and Lowe, D.G. 1997. Shape indexing using approximate nearest-neighbour search in high dimensional spaces. In Conference on Computer Vision and Pattern Recognition, Puerto Rico, pp. 1000-1006.
[15] Arya, S., and Mount, D.M. 1993. Approximate nearest neighbor queries in fixed dimensions. In Fourth Annual ACM-SIAM Symposium on Discrete Algorithms (SODA’93), pp. 271-280.
[16] Arya, S., Mount, D.M., Netanyahu, N.S., Silverman, R., and Wu, A.Y. 1998. An optimal algorithm for approximate nearest neighbor searching. Journal of the ACM, 45:891-923.
[17] M. A. Fischler and R. C. Bolles. Random sample consensus: A paradigm for model fittingith applications to image analysis and automated cartography. Communications of the ACM, 24:381–395, 1981.
[18] P. J. Huber. Robust Statistics. Wiley, 1981.
[19] P. J. Rousseeuw and A. M. Leroy. Robust Regression and Outlier Detection. Wiley, 1987.
[20] J. Mendel. Lessons in Estimation Theory for Signal Processing, Communication and Control. Prentice-Hall, Englewood-Cliffs, 1995.
[21] Z. Zhang. Parameter estimation techniques: A tutorial with application to conic fitting. Image and Vision Computing Journal, 25(1):59–76, 1997.
[22] C. V. Stewart. Robust parameter estimation in computer vision. SIAM Review, 41(3):513–537, September 1999.
[23] C. V. Stewart. MINPRAN: A new robust estimator for computer vision. IEEE Transactions on Pattern Analysis and Machine Intelligence, 17(10):925–938, October 1995.
[24] R.Hartley and A.Zisserman. Multiple View Geometry in Computer Vision. Cambridge University Press, second edition, 2003.
[25] B. J. Tordoff and D. W. Murray. Guided-MLESAC: Faster image transform estimation by using matching priors. IEEE Transactions on Pattern Analysis and Machine Intelligence, 27(10):1523–1535, October 2005.
[26] P. Burt and E. Adelson. A Multiresolution spline with Application to Image Mosaics. ACM Transactions on Graphics, Vol. 2, No. 4, October 1983. Pg. 217-236.