基于模糊集理论的图像增强算法研究
|
摘要
图像增强作为一种低层数字图象处理手段,往往具有模糊性。本文简要介绍了模糊理论的发展以及基于模糊理论的图象增强的发展及现状,并对现有基于模糊集合理论的图象增强算法,主要是对隶属函数和模糊增强算子(INT)进行了分析比较。首先,为了能根据不同图像的输入特性实现自适应的模糊增强,在分析已有模糊熵定义的基础上,基于信息增加的指数特性,提出了一种新的模糊熵定义:在最大熵准则下,利用枚举法实现了图像的自适应增强,实验结果表明自适应增强效果在主客观评价上都较好,且运行时间比其他模糊熵定义下的自适应增强算法时间短。其次,为了使自适应模糊增强算法接近实时应用,利用遗传算法的全局寻优性能和并行性,改进了基于枚举法的自适应模糊增强算法,将新的模糊熵定义作为适应度函数,在保证处理结果主客观评价不变的条件下,处理速度有了大幅提高;采用遗传算法后,与传统优化算法相比,自适应增强算法的参数优化部分的时间减少了60%。最后,利用广义模糊增强的思想以及考虑图像的物体区域、背景区域以及边缘区域的特点提出了一种更快速、边缘检测效果更好的模糊增强算子,尤其对低对比度图像、图像中含有较精细的部分以及纹理丰富的图像检测效果较好,且时间有所减少。
As a means of low level image processing, image enhancement usually has ambiguity and vagueness. The development of fuzzy theory and the status quo of image enhancement based on fuzzy theory are briefly introduced. And the image enhancement algorithms in existence based on fuzzy set theory, esp. the fuzzy membership functions and the INT's, are analyzed and compared. Firstly, a novel fuzzy entropy definition is proposed based on the analyzing the existed fuzzy entropy definitions and the exponential behavior of information-gain. Self-adaptive image enhancement is achieved with the novel fuzzy entropy definition and the maximum entropy principle using the enumerating algorithm. And the experiment results show that the effect of the self-adaptive image enhancement is good both in subject evaluation and object evaluation and that its processing time is shorter than that of other algorithms based on the existed fuzzy entropy definitions. Secondly, the self-adaptive enhancement algorithm based on enumerating algorithm is improved with the help of the searching optimum capacity in wholeness and the parallelism of genetic algorithm in order to save processing time and close with its real time applications. And the new fuzzy entropy definition is taken as fitness function. Its experiment results show that the subject evaluation and object evaluation are not changed and that the processing time of the algorithm's optimum part is shorter 60% than that of enumerating algorithm. At last, a fuzzy enhancement operator which can detect images' edges faster and better than GFOs is proposed based on the general fuzzy enhancement and the properties of the object region, background region and edges region of images. And the experiment results show that the algorithm is especially fit for low contrast images and images including fine parts and much texture and that processing time is cut down a little.
As a means of low level image processing, image enhancement usually has ambiguity and vagueness. The development of fuzzy theory and the status quo of image enhancement based on fuzzy theory are briefly introduced. And the image enhancement algorithms in existence based on fuzzy set theory, esp. the fuzzy membership functions and the INT's, are analyzed and compared. Firstly, a novel fuzzy entropy definition is proposed based on the analyzing the existed fuzzy entropy definitions and the exponential behavior of information-gain. Self-adaptive image enhancement is achieved with the novel fuzzy entropy definition and the maximum entropy principle using the enumerating algorithm. And the experiment results show that the effect of the self-adaptive image enhancement is good both in subject evaluation and object evaluation and that its processing time is shorter than that of other algorithms based on the existed fuzzy entropy definitions. Secondly, the self-adaptive enhancement algorithm based on enumerating algorithm is improved with the help of the searching optimum capacity in wholeness and the parallelism of genetic algorithm in order to save processing time and close with its real time applications. And the new fuzzy entropy definition is taken as fitness function. Its experiment results show that the subject evaluation and object evaluation are not changed and that the processing time of the algorithm's optimum part is shorter 60% than that of enumerating algorithm. At last, a fuzzy enhancement operator which can detect images' edges faster and better than GFOs is proposed based on the general fuzzy enhancement and the properties of the object region, background region and edges region of images. And the experiment results show that the algorithm is especially fit for low contrast images and images including fine parts and much texture and that processing time is cut down a little.
引文
[1] Hua Li, Hyun S. Yang. Fast and reliable image enhancement using fuzzy relaxation technique. IEEE Transactions on Systems, Man and Cybernetics, 1989,19(5):1276～1281
[2] 闻新等.MATLAB模糊逻辑工具箱的分析与应用.北京:科学出版社,2001
[3] http://watfor.uwaterloo.ca/tizhoosh/history.htm
[4] Osama K. Alshaykh, Srikrishna Ramaswamy, Hsien-Sen Hung. Fuzzy techniques for image enhancement and reconstruction. Second IEEE International Conference on Fuzzy Systems, 1993,1:582～587
[5] Sankar K. Pal. Fuzzy sets in image processing and recognition. IEEE International Conference on Fuzzy Systems, 1992:119～126
[6] 徐立中.数字图像的智能信息处理.北京:国防工业出版社,2001.41～47,63～93
[7] 陈武凡,鲁贤庆,陈建军等.彩色图象边缘检测的新算法——广义模糊算子法.中国科学(A辑),1995,25(2):219～224
[8] 靳蕃.神经计算智能基础原理·方法.成都:西南交通大学出版社,2000.177～215
[9] 赵振宇,徐用懋.模糊理论和神经网络的基础与应用.北京:清华大学出版社,1996.5～25
[10] 安良.模糊理论及其在图像分割中的应用研究:[硕士学位论文].合肥:合肥工业大学,2003
[11] http://watfor.uwaterloo.ca/tizhoosh/wfip.htm
[12] Dong-ling Peng, Tie-jun Wu. A generalized image enhancement algorithm using fuzzy sets and its application. Proceedings of the 1st International Conference on Machine Learning and Cybemetics,2002:820～823
[13] H.R. Tizhoosh, G. Krell, B. Michaelis. On fuzzy enhancement of megavoltage images in radiation therapy. Proceedings of the Sixth IEEE International Conference on Fuzzy Systems, 1997,3:1398～1404
[14] S. K. Pal, R. A. King. Image enhancement using smoothing with fuzzy sets.IEEE Transactions on Systems, Man and Cybernetics, 1981,11(7):494～501
[15] S. K. Pal, R. A. King. On edge detection of X-ray images using fuzzy sets. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1983,5(1):69～77
[16] Zheng Chunhong, Jiao Licheng, Chen Xiping, Yuan Zhanting. A new fast fuzzy processing method for B-scan image. Proceedings of the 4th World Congress on Intelligent Control and Automation,2002:6～9
[17] S. K. Pal, R. A. King. Histogram equalization with S and Pi functions in detecting X-ray edges. Electronics Letters, 1981,17(8):302～304
[18] Hua Li, Hyun S. Yang. Fast and reliable image enhancement using fuzzy relaxation technique. IEEE Transactions on Systems, Man and Cybernetics, 1989,19(5):1276～1281
[19] Fa-Shen Leou, Kuei-Ann Wen. Image enhancement based on the visual model using the concept of fuzzy set. IEEE International Symposium on Circuits and Systems, 1992,5:2581～2584
[20] Tito G. Amaral, Manuel M. Crisostomo, A. Traca de Almeida. Image thresholding by minimisation of fuzzy compactness and linear index of fuzziness. Proceedings of IEEE International Conference on Fuzzy Systems, 1999,2:1116～1121
[21] H.R. Tizhoosh, G. Krell, B. Michaelis. λ-enhancement: contrast adaptation based on optimization of image fuzziness. Fuzzy Systems Proceedings of IEEE World Congress on Computational Intelligence, 1998,2:1548～1553
[22] C.C. Leung, F.H.Y. Chan, K.Y. Lam, P.C.K. Kwok, W.F. Chen. Thyroid cancer cells boundary location by a fuzzy edge detection method. International Conference on Pattern Recognition,2000,4:360～363
[23] 王晖,张基宏.图像边缘检测的区域对比度模糊增强算法.电子学报,2000,28(1):45～47
[24] 王晖.区域对比度模糊增强及其在医学图象边界检测中的应用.中国生物医学工程学报,2000,19(3):353～355
[25] 刘恒殊,黄廉卿.基于模糊集理论的医学CR图像增强.光学精密工程,2002,10(1):94～97
[26] 韩培友,郝重阳,董桂云.基于GFS的双线性快速模糊增强图像边界检测新算法.计算机工程,2004,30(19):6～7
[27] 张家树,李明华.一种新的图像模糊增强技术.西南师范大学学报(自然科学版),1994,19(1):37～41
[28] 冷寒冰,拜丽萍,刘上乾.基于模糊域的红外图像自适应降噪增强技术.红外与激光工程,2003,32(5):521～522
[29] 张世华,宋振明.一种基于模糊增强的图像边缘提取改进算法.株洲工学院学报,2003,17(5):45～48
[30] 张弘,庄天戈.一种用于胸部CT的改进模糊图像增强算法.上海交通大学学报,2002,36(7):1026～1028
[31] C.C. Leung, P.C.K. Kwok, F.H.Y. Chan, W.K. Tsui. Normalization of contrast in document images using generalized fuzzy operator with least square method. Proceedings of 16th International Conference on Pattern Recognition,2002,3:115～118
[32] 任彬,汪炳权,黄勇.医学B超图象的模糊增强处理方法.中国图象图形学报,1997,1(2,3):133～136
[33] Pan Rongjiang, Meng Xiangxu. A method of local enhancement based on fuzzy set theory. Proceedings of 3rd World Congress on Intelligent Control and Automation,2000:1751～1753
[34] 张运杰,于东.改进的模糊增强算法.辽宁工程技术大学学报(自然科学版),2001,20(5):610～620
[35] Wen-Bing Tao, Jin-Wen Tian, Jian Liu. Image segmentation by three-level thresholding based on maximum fuzzy entropy and genetic algorithm. Pattern Recognition Letters,2003,24:3069～3078
[36] Li XiangJi, Ding RunTao. Fuzzy morphological operators to edge enhancing of images. Proceedings of International Conference on Signal Processing, 1998,2:1017～1020
[37] 李红松,侯朝桢.一种新的模糊边缘检测算法.计算机工程,2003,29(9):1～2
[38] 王英健,史绍强,唐贤瑛.基于模糊粗糙集理论的视觉特性图像增强.长沙理工大学学报(自然科学版),2004,1(2):79～83
[39] Fa-Shen Leou, Kuei-Ann Wen. Image enhancement based on the visual model using the concept of fuzzy set. Proceedings of IEEE International Symposium on Circuits and Systems, 1992,5:2581～2584
[40] 刘金华,唐竞新,龙图景.一种改进的模糊边缘检测快速算法.系统仿真学报,2003,15(2):273～274
[41] 孙伟,夏良正,潘泓.一种基于模糊划分的边缘检测算法.中国图象图形学报,2004,9(1):18～22
[42] http://en.wikipedia.org/wiki/Principle_of_maximum_entropy
[43] H.D. Cheng, Yen-Hung Chen, Ying Sun. A novel fuzzy entropy approach to image enhancement and thresholding. Signal Processing, 1999,75:277-301
[44] N.R. Pal, S.K. Pal. Object-background segmentation using new definitions of entropy. IEE Proceedings-E, 1989,136(4):284～295,
[45] 汪亚明,吴国忠,赵匀等.基于模糊熵及遗传算法的图像增强技术.农业机械学报,2003,34(3):96～98
[46] 周明,孙树栋.遗传算法原理及应用.北京:国防工业出版社,1999:18～64
[47] 王小平,曹立明.遗传算法——理论、应用与软件实现.西安:西安交通大学出版社,2002:7～49
[2] 闻新等.MATLAB模糊逻辑工具箱的分析与应用.北京:科学出版社,2001
[3] http://watfor.uwaterloo.ca/tizhoosh/history.htm
[4] Osama K. Alshaykh, Srikrishna Ramaswamy, Hsien-Sen Hung. Fuzzy techniques for image enhancement and reconstruction. Second IEEE International Conference on Fuzzy Systems, 1993,1:582～587
[5] Sankar K. Pal. Fuzzy sets in image processing and recognition. IEEE International Conference on Fuzzy Systems, 1992:119～126
[6] 徐立中.数字图像的智能信息处理.北京:国防工业出版社,2001.41～47,63～93
[7] 陈武凡,鲁贤庆,陈建军等.彩色图象边缘检测的新算法——广义模糊算子法.中国科学(A辑),1995,25(2):219～224
[8] 靳蕃.神经计算智能基础原理·方法.成都:西南交通大学出版社,2000.177～215
[9] 赵振宇,徐用懋.模糊理论和神经网络的基础与应用.北京:清华大学出版社,1996.5～25
[10] 安良.模糊理论及其在图像分割中的应用研究:[硕士学位论文].合肥:合肥工业大学,2003
[11] http://watfor.uwaterloo.ca/tizhoosh/wfip.htm
[12] Dong-ling Peng, Tie-jun Wu. A generalized image enhancement algorithm using fuzzy sets and its application. Proceedings of the 1st International Conference on Machine Learning and Cybemetics,2002:820～823
[13] H.R. Tizhoosh, G. Krell, B. Michaelis. On fuzzy enhancement of megavoltage images in radiation therapy. Proceedings of the Sixth IEEE International Conference on Fuzzy Systems, 1997,3:1398～1404
[14] S. K. Pal, R. A. King. Image enhancement using smoothing with fuzzy sets.IEEE Transactions on Systems, Man and Cybernetics, 1981,11(7):494～501
[15] S. K. Pal, R. A. King. On edge detection of X-ray images using fuzzy sets. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1983,5(1):69～77
[16] Zheng Chunhong, Jiao Licheng, Chen Xiping, Yuan Zhanting. A new fast fuzzy processing method for B-scan image. Proceedings of the 4th World Congress on Intelligent Control and Automation,2002:6～9
[17] S. K. Pal, R. A. King. Histogram equalization with S and Pi functions in detecting X-ray edges. Electronics Letters, 1981,17(8):302～304
[18] Hua Li, Hyun S. Yang. Fast and reliable image enhancement using fuzzy relaxation technique. IEEE Transactions on Systems, Man and Cybernetics, 1989,19(5):1276～1281
[19] Fa-Shen Leou, Kuei-Ann Wen. Image enhancement based on the visual model using the concept of fuzzy set. IEEE International Symposium on Circuits and Systems, 1992,5:2581～2584
[20] Tito G. Amaral, Manuel M. Crisostomo, A. Traca de Almeida. Image thresholding by minimisation of fuzzy compactness and linear index of fuzziness. Proceedings of IEEE International Conference on Fuzzy Systems, 1999,2:1116～1121
[21] H.R. Tizhoosh, G. Krell, B. Michaelis. λ-enhancement: contrast adaptation based on optimization of image fuzziness. Fuzzy Systems Proceedings of IEEE World Congress on Computational Intelligence, 1998,2:1548～1553
[22] C.C. Leung, F.H.Y. Chan, K.Y. Lam, P.C.K. Kwok, W.F. Chen. Thyroid cancer cells boundary location by a fuzzy edge detection method. International Conference on Pattern Recognition,2000,4:360～363
[23] 王晖,张基宏.图像边缘检测的区域对比度模糊增强算法.电子学报,2000,28(1):45～47
[24] 王晖.区域对比度模糊增强及其在医学图象边界检测中的应用.中国生物医学工程学报,2000,19(3):353～355
[25] 刘恒殊,黄廉卿.基于模糊集理论的医学CR图像增强.光学精密工程,2002,10(1):94～97
[26] 韩培友,郝重阳,董桂云.基于GFS的双线性快速模糊增强图像边界检测新算法.计算机工程,2004,30(19):6～7
[27] 张家树,李明华.一种新的图像模糊增强技术.西南师范大学学报(自然科学版),1994,19(1):37～41
[28] 冷寒冰,拜丽萍,刘上乾.基于模糊域的红外图像自适应降噪增强技术.红外与激光工程,2003,32(5):521～522
[29] 张世华,宋振明.一种基于模糊增强的图像边缘提取改进算法.株洲工学院学报,2003,17(5):45～48
[30] 张弘,庄天戈.一种用于胸部CT的改进模糊图像增强算法.上海交通大学学报,2002,36(7):1026～1028
[31] C.C. Leung, P.C.K. Kwok, F.H.Y. Chan, W.K. Tsui. Normalization of contrast in document images using generalized fuzzy operator with least square method. Proceedings of 16th International Conference on Pattern Recognition,2002,3:115～118
[32] 任彬,汪炳权,黄勇.医学B超图象的模糊增强处理方法.中国图象图形学报,1997,1(2,3):133～136
[33] Pan Rongjiang, Meng Xiangxu. A method of local enhancement based on fuzzy set theory. Proceedings of 3rd World Congress on Intelligent Control and Automation,2000:1751～1753
[34] 张运杰,于东.改进的模糊增强算法.辽宁工程技术大学学报(自然科学版),2001,20(5):610～620
[35] Wen-Bing Tao, Jin-Wen Tian, Jian Liu. Image segmentation by three-level thresholding based on maximum fuzzy entropy and genetic algorithm. Pattern Recognition Letters,2003,24:3069～3078
[36] Li XiangJi, Ding RunTao. Fuzzy morphological operators to edge enhancing of images. Proceedings of International Conference on Signal Processing, 1998,2:1017～1020
[37] 李红松,侯朝桢.一种新的模糊边缘检测算法.计算机工程,2003,29(9):1～2
[38] 王英健,史绍强,唐贤瑛.基于模糊粗糙集理论的视觉特性图像增强.长沙理工大学学报(自然科学版),2004,1(2):79～83
[39] Fa-Shen Leou, Kuei-Ann Wen. Image enhancement based on the visual model using the concept of fuzzy set. Proceedings of IEEE International Symposium on Circuits and Systems, 1992,5:2581～2584
[40] 刘金华,唐竞新,龙图景.一种改进的模糊边缘检测快速算法.系统仿真学报,2003,15(2):273～274
[41] 孙伟,夏良正,潘泓.一种基于模糊划分的边缘检测算法.中国图象图形学报,2004,9(1):18～22
[42] http://en.wikipedia.org/wiki/Principle_of_maximum_entropy
[43] H.D. Cheng, Yen-Hung Chen, Ying Sun. A novel fuzzy entropy approach to image enhancement and thresholding. Signal Processing, 1999,75:277-301
[44] N.R. Pal, S.K. Pal. Object-background segmentation using new definitions of entropy. IEE Proceedings-E, 1989,136(4):284～295,
[45] 汪亚明,吴国忠,赵匀等.基于模糊熵及遗传算法的图像增强技术.农业机械学报,2003,34(3):96～98
[46] 周明,孙树栋.遗传算法原理及应用.北京:国防工业出版社,1999:18～64
[47] 王小平,曹立明.遗传算法——理论、应用与软件实现.西安:西安交通大学出版社,2002:7～49