遥感影像中水上桥梁目标的识别方法研究
|
摘要
遥感图像目标自动识别技术是计算机模式识别与图像处理领域非常活跃的研究课题。自然景物中的桥梁目标是一种典型人造目标，通过计算机视觉技术对其进行自动识别的研究，不论在军用还是民用上都有重大的意义。
本文的工作主要是对SPOT遥感影像中水上桥梁目标进行识别，即以计算机为手段，从复杂的遥感图像中检测出水上桥梁，并对其精确定位。
迄今为止，目标识别算法开发的策略基本上可以归纳为两种：一种称为由下而上的数据驱动型策略，另一种称为由上而下的知识(假设)驱动型策略。本文所提出的识别算法采用的是知识驱动型的策略，算法分低、中、高三个处理层次，各个层次、各个步骤都有相应的知识基进行引导。
鉴于桥梁目标在本文使用的两类不同分辨率的SPOT影像(10米／象素的中高分辨率影像和20米／象素的低分辨率)中的表现有较大的区别，因此本文在“知识驱动”这个总体策略不变的前提下，根据目标在两类影像中的不同特点分别提出了两套不同的识别算法，在低、中、高各层处理中都采用了不同的处理方法。
本文提出的识别方法与传统的方法相比，通用性强，检测速度快，较好地实现了水上桥梁目标识别的自动化。
The technology of object automatic recognition in remote sensing image is a very active research field in recent years. Bridge is a typical object in the nature, the research of the automatic recognition of bridge has great meaning to both the military and the civil.
Our work is to recognize the object of bridges which are across the rivers in the SPOT remote sensing images, i.e., to automatically find out bridges in complex remote sensing image by using technology of computer vision.
To this day, there are two kinds of strategies in the technology of object automatic recognition: one is data driven, and the other is knowledge driven. The algorithm of bridge recognition in this paper adopts the knowledge driven strategy.
It is made up of three processing steps--the primary step, intermediate step and
advanced step. Each step will be guided by its corresponding knowledge.
Considering the different representation of bridges in the two different kinds of remote sensing images in different resolution(one is 10 meters/pixel , and the other is 20 meters/pixel), this paper respectively presents two suits of algorithms for these two kinds of images. Each step in these two different algorithms will take different processing method, yet they are all driven by knowledge.
The method purposed in this paper is much faster and also more general than traditional recognition methods, it does well in realizing the automatic recognition of bridges across the rivers in remote sensing image.
本文的工作主要是对SPOT遥感影像中水上桥梁目标进行识别，即以计算机为手段，从复杂的遥感图像中检测出水上桥梁，并对其精确定位。
迄今为止，目标识别算法开发的策略基本上可以归纳为两种：一种称为由下而上的数据驱动型策略，另一种称为由上而下的知识(假设)驱动型策略。本文所提出的识别算法采用的是知识驱动型的策略，算法分低、中、高三个处理层次，各个层次、各个步骤都有相应的知识基进行引导。
鉴于桥梁目标在本文使用的两类不同分辨率的SPOT影像(10米／象素的中高分辨率影像和20米／象素的低分辨率)中的表现有较大的区别，因此本文在“知识驱动”这个总体策略不变的前提下，根据目标在两类影像中的不同特点分别提出了两套不同的识别算法，在低、中、高各层处理中都采用了不同的处理方法。
本文提出的识别方法与传统的方法相比，通用性强，检测速度快，较好地实现了水上桥梁目标识别的自动化。
The technology of object automatic recognition in remote sensing image is a very active research field in recent years. Bridge is a typical object in the nature, the research of the automatic recognition of bridge has great meaning to both the military and the civil.
Our work is to recognize the object of bridges which are across the rivers in the SPOT remote sensing images, i.e., to automatically find out bridges in complex remote sensing image by using technology of computer vision.
To this day, there are two kinds of strategies in the technology of object automatic recognition: one is data driven, and the other is knowledge driven. The algorithm of bridge recognition in this paper adopts the knowledge driven strategy.
It is made up of three processing steps--the primary step, intermediate step and
advanced step. Each step will be guided by its corresponding knowledge.
Considering the different representation of bridges in the two different kinds of remote sensing images in different resolution(one is 10 meters/pixel , and the other is 20 meters/pixel), this paper respectively presents two suits of algorithms for these two kinds of images. Each step in these two different algorithms will take different processing method, yet they are all driven by knowledge.
The method purposed in this paper is much faster and also more general than traditional recognition methods, it does well in realizing the automatic recognition of bridges across the rivers in remote sensing image.
引文
[1] 夏德深，傅德胜．现代图像处理技术和应用．东南大学出版社，1997．
[2] Baker D.C. Hwang S.S. Aggarwal J.K. Detection and Segmentation of Man-Made Objects in Outdoor Scenes:Concret Bridge. J.Opt.Soc Am 1989(6).938-950.
[3] Baker D.C. Aggarwal J.K. Hwang S.S. Geometry Guided Segmentation of Outdoor Scenes. SPIE App of AI, 1988(9317):576-583.
[4] 徐胜荣，李忠兴．自然景物中桥梁目标识别方法的研究．浙江大学学报，1995(5)：275-281．
[5] 徐胜荣，徐凡，叶秀清，荆仁杰．基于知识的桥梁目标识别方法研究．模式识别与人工智能，1992，5(2)：123-128．
[6] 姜骊黎，史册，杨海波，姚庆栋．遥感图象中水上桥梁的识别．模式识别与人工智能，2000，13(2)：214-217．
[7] 王伟明，徐胜荣，姚庆栋，唐慧明．一种使用黑板模型实现的图象分割策略及在桥梁目标识别中的应用．计算机学报，1995，18(2)：127-134．
[8] 左震，张天序，汪国有．远距红外图象中桥梁目标识别方法研究．电子学报，1998，26(11)：6-10．
[9] HouBiao, LiYing, JiaoLicheng. Segmentation and Recognition of Bridges in High Resolution SAR Images.
[10] Daniel Crevier, Richard Lepage. Knowledge-based Image Understanding Systems: A Survey, COMPUTER VISION AND IMAGE UNDERSTANDING, Vol. 67, No. 2, August, pp. 161-185, 1997, ARTICLE NO. IV960520
[11] 艾海舟，武勃．图像处理、分析与机器视觉．北京，人民邮电出版社，2003
[12] 章毓晋．图像处理和分析，北京：清华大学出版社，1999
[13] 陈武凡．小波分析及其在图像处理中的应用．北京：科学出版社，2002．
[14] Lu Guanming. Region growth watershed algorithm and its applications in image sequence segmentation [J]. Journal Nanjing University of Posts and Communications, 2000, 20(3): 51-54(in Chinese)
[15] 徐建华．图像处理与分析，北京：科学出版社，1992
[16] 朱立新，一种基于双峰直方图谷底检测的阈值新方法
[17] 梁华为．直接从双峰直方图确定二值化阈值．模式识别与人工智能，2002，15(2)：253-255
[18] 李军，周月琴．小波变换用于影像分割的研究．中国图象图形学报，1997，2(4)：213-219
[19] 任彬，汪炳权，罗斌．基于直方图指数平滑的阈值和峰点自动检测方法．中国图象图形学报，1997，2(4)：230-233
[20] 史册，徐胜荣，荆仁杰，姚庆栋．基于团块的特征提取．计算机学报，1997，20(12)：1124-1128．
[21] 孙兆林．Matlab 6.x图像处理，北京，清华大学出版社，2002
[22] Milan Sonka, Vaclav Hlavac, Roger Boyle. Image Processing, Analysis, and Machine Version. Thomson Learning and PT Press. 1999
[23] Maragos. P., Tutorial on advances in morphological image processing and analysis [J]. Optical Engineering, 1987, 26(7): 623-632
[24] Wegman E. J., Nonparametric probability density estimation, A summary of available methods, Techno metrics[M] , 1986.
[25] Serra. J., Image Analysis and mathematical morphology[M] , 1982.
[26] Hatalick R, Zhuang X. Image analysis using mathematical morphology [J].IEEE Trans. On Patten Analysis and Machine Intelligence,1987, 9(4): 532-550
[27] Lu Guanming. Region growth watershed algorithm and its applications in image sequence segmentation [J]. Joumal Nanjing University of Posts and Communications, 2000, 20(3): 51-54(in Chinese)
[28] Vincent. L, Soille, P. Watershed in digital spaces: An efficient algorithm based immersion simulations[j]. IEEE Trans on Pattern Analysis & Machine Intelligence, 1991, 13(6): 583-598.
[29] 杨静宇，曹雨龙．计算机图像处理及常用算法手册．南京大学出版社．南京．1997
[30] 王晖，张基宏．图像边界检测的区域对比度模糊增强算法．电子学报，2000，28(1)：45-47
[31] Pal S K, King R A. Image Enhancement Using Fuzzy Sets. Electronics Letters, 1980, 16(10):376～378
[32] Pal S K, King R A. Image Enhancement Using Smoothing with Fuzzy Sets. IEEE Trans on Systems, Man, and Cybernetics, 1981, 11(7):494～501
[33] Borgi A, Akdag H. Knowledge Based Supervised Fuzzy-Classification:An Application to Image Processing. Annals of Mathematics and Artificial Intelligence, 2001, 32(1):67～86
[34] Cheng H D, Cheng Y H, Sun Y. A Novel Fuzzy Entropy Approach to Image Enhancement and Thresholding. Singal Processing, 1999, 75(3):277～301
[35] Acton S T. On Fuzzy Nonlinear Regression for Image Enhancement. Journal of Mathematical Imaging and Vision, 1998, 8(3):239～253
[36] 彭冬亮，吴铁军．一种广义的图像模糊增强算法．模式识别与人工智能，2003，16(3)：338-341
[37] 陈武凡，谢新鹏，洪文松．广义模糊集合论及其在图像处理中的新应用．中国生物医学工程学报．1999，18(1)：66～73
[38] 洪文松，陈武凡．广义模糊集合论及其应用模式．暨南大学学报，2000，21(1)：77～79
[39] 谢新鹏，陈武凡．一种基于广义模糊性质集的图像分割方法．
[40] 周杰，陈武夫．广义模糊算子应用于图像边缘检测中的参数选择与经验计算．中国图像图形学报．1997，2(2，3)：108～112．
[41] L. A. Zadeh, Fuzzy Sets, Information and Control, 1965. vol. 8:338～353
[42] 陈武凡．彩色图像边缘检测的新算法．中国科学(A辑)，1995，25(2)：219～224
[2] Baker D.C. Hwang S.S. Aggarwal J.K. Detection and Segmentation of Man-Made Objects in Outdoor Scenes:Concret Bridge. J.Opt.Soc Am 1989(6).938-950.
[3] Baker D.C. Aggarwal J.K. Hwang S.S. Geometry Guided Segmentation of Outdoor Scenes. SPIE App of AI, 1988(9317):576-583.
[4] 徐胜荣，李忠兴．自然景物中桥梁目标识别方法的研究．浙江大学学报，1995(5)：275-281．
[5] 徐胜荣，徐凡，叶秀清，荆仁杰．基于知识的桥梁目标识别方法研究．模式识别与人工智能，1992，5(2)：123-128．
[6] 姜骊黎，史册，杨海波，姚庆栋．遥感图象中水上桥梁的识别．模式识别与人工智能，2000，13(2)：214-217．
[7] 王伟明，徐胜荣，姚庆栋，唐慧明．一种使用黑板模型实现的图象分割策略及在桥梁目标识别中的应用．计算机学报，1995，18(2)：127-134．
[8] 左震，张天序，汪国有．远距红外图象中桥梁目标识别方法研究．电子学报，1998，26(11)：6-10．
[9] HouBiao, LiYing, JiaoLicheng. Segmentation and Recognition of Bridges in High Resolution SAR Images.
[10] Daniel Crevier, Richard Lepage. Knowledge-based Image Understanding Systems: A Survey, COMPUTER VISION AND IMAGE UNDERSTANDING, Vol. 67, No. 2, August, pp. 161-185, 1997, ARTICLE NO. IV960520
[11] 艾海舟，武勃．图像处理、分析与机器视觉．北京，人民邮电出版社，2003
[12] 章毓晋．图像处理和分析，北京：清华大学出版社，1999
[13] 陈武凡．小波分析及其在图像处理中的应用．北京：科学出版社，2002．
[14] Lu Guanming. Region growth watershed algorithm and its applications in image sequence segmentation [J]. Journal Nanjing University of Posts and Communications, 2000, 20(3): 51-54(in Chinese)
[15] 徐建华．图像处理与分析，北京：科学出版社，1992
[16] 朱立新，一种基于双峰直方图谷底检测的阈值新方法
[17] 梁华为．直接从双峰直方图确定二值化阈值．模式识别与人工智能，2002，15(2)：253-255
[18] 李军，周月琴．小波变换用于影像分割的研究．中国图象图形学报，1997，2(4)：213-219
[19] 任彬，汪炳权，罗斌．基于直方图指数平滑的阈值和峰点自动检测方法．中国图象图形学报，1997，2(4)：230-233
[20] 史册，徐胜荣，荆仁杰，姚庆栋．基于团块的特征提取．计算机学报，1997，20(12)：1124-1128．
[21] 孙兆林．Matlab 6.x图像处理，北京，清华大学出版社，2002
[22] Milan Sonka, Vaclav Hlavac, Roger Boyle. Image Processing, Analysis, and Machine Version. Thomson Learning and PT Press. 1999
[23] Maragos. P., Tutorial on advances in morphological image processing and analysis [J]. Optical Engineering, 1987, 26(7): 623-632
[24] Wegman E. J., Nonparametric probability density estimation, A summary of available methods, Techno metrics[M] , 1986.
[25] Serra. J., Image Analysis and mathematical morphology[M] , 1982.
[26] Hatalick R, Zhuang X. Image analysis using mathematical morphology [J].IEEE Trans. On Patten Analysis and Machine Intelligence,1987, 9(4): 532-550
[27] Lu Guanming. Region growth watershed algorithm and its applications in image sequence segmentation [J]. Joumal Nanjing University of Posts and Communications, 2000, 20(3): 51-54(in Chinese)
[28] Vincent. L, Soille, P. Watershed in digital spaces: An efficient algorithm based immersion simulations[j]. IEEE Trans on Pattern Analysis & Machine Intelligence, 1991, 13(6): 583-598.
[29] 杨静宇，曹雨龙．计算机图像处理及常用算法手册．南京大学出版社．南京．1997
[30] 王晖，张基宏．图像边界检测的区域对比度模糊增强算法．电子学报，2000，28(1)：45-47
[31] Pal S K, King R A. Image Enhancement Using Fuzzy Sets. Electronics Letters, 1980, 16(10):376～378
[32] Pal S K, King R A. Image Enhancement Using Smoothing with Fuzzy Sets. IEEE Trans on Systems, Man, and Cybernetics, 1981, 11(7):494～501
[33] Borgi A, Akdag H. Knowledge Based Supervised Fuzzy-Classification:An Application to Image Processing. Annals of Mathematics and Artificial Intelligence, 2001, 32(1):67～86
[34] Cheng H D, Cheng Y H, Sun Y. A Novel Fuzzy Entropy Approach to Image Enhancement and Thresholding. Singal Processing, 1999, 75(3):277～301
[35] Acton S T. On Fuzzy Nonlinear Regression for Image Enhancement. Journal of Mathematical Imaging and Vision, 1998, 8(3):239～253
[36] 彭冬亮，吴铁军．一种广义的图像模糊增强算法．模式识别与人工智能，2003，16(3)：338-341
[37] 陈武凡，谢新鹏，洪文松．广义模糊集合论及其在图像处理中的新应用．中国生物医学工程学报．1999，18(1)：66～73
[38] 洪文松，陈武凡．广义模糊集合论及其应用模式．暨南大学学报，2000，21(1)：77～79
[39] 谢新鹏，陈武凡．一种基于广义模糊性质集的图像分割方法．
[40] 周杰，陈武夫．广义模糊算子应用于图像边缘检测中的参数选择与经验计算．中国图像图形学报．1997，2(2，3)：108～112．
[41] L. A. Zadeh, Fuzzy Sets, Information and Control, 1965. vol. 8:338～353
[42] 陈武凡．彩色图像边缘检测的新算法．中国科学(A辑)，1995，25(2)：219～224