基于边缘提取与颜色目标定位的图像检索算法
|
摘要
基于内容的图像检索是图像检索技术发展的一个重要方向,是管理海量数字图像和视频信息的一种有效手段,其本质是利用图像特有的理解方式来实现自动查找,实现的是一对多的相似查找。
本文重点研究以图像边缘特征和颜色特征为模型的检索算法,主要内容包括:首先介绍基于内容的图像检索技术,总结了颜色特征、纹理特征和形状特征的表达方式和提取方法,并对图像检索预处理的图像分割与像素分类作了研究,分析了现有检索算法的局限性与不足之处。
提出了两种综合利用图像边缘结构特征与颜色特征的图像检索算法。(1)基于Canny算子提取图像边缘与颜色目标定位的图像检索算法。(2)基于小波域提取图像边缘与颜色目标定位的图像检索算法。这两种算法分别利用Canny空间算子与小波变换域两种边缘提取方法对图像的形状特征进行描述,在待检索图像库中查找到与示例图像边缘相关度大的图像,构成一个新的检索图像库,然后再利用主颜色划分图像目标,利用有效的目标部分进行相似匹配,完成对示例图像的检索,仿真实验证明,这种结构的检索算法可以有效捕捉边缘信息,提高检索性能。
本文对边缘的提取不作为目标分割的手段,仅视为反映图像几何结构分布的低层次特征,而且引入小波域变换,避免了因图像理解的目标分割上的困难,有效地反映了图像的结构分布,提高了图像检索抗干扰性。
Content-Based Image Retrieval (CBIR) is an important research field in the development of image retrieval. It is an effective method of administering magnanimity digital image and video information. Its essence is to realize the image automatic search utilizing the special understanding way of the image. It can be used to realize one-to-many similar search.
This paper mainly studies retrieval algorithm based on image edge and color features. The main contents include:
First of all, the technology of CBIR was simply introduced. A great deal of methods of extraction and expression about the color, texture and shape features were summed up and summarized in detail. Image segmentation and pixels classifying of image retrieval pretreatment was studied. The limitation and deficiency of the existing retrieval algorithm was analyzed.
Second, two image retrieval algorithms using image edge structure and color features were put forward. (1) Image Retrieval Based on Canny Edge Detection and Goal Location. (2) Image Retrieval Based on the Wavelet Transform Edge Detection and Goal Location. The edge feature was described and extracted in the two kinds of domains: the spatial domain and the transform domain. The images that have the close correlation edge to the example image was searched to form a new multi-images database. Then it used the dominant color to divide image target, used effective target part for similarity matching and finishing the retrieval of the example images. Experiments show that this kind of arithmetic can well and efficiently detect the edge feature and improve the efficiency.
The edge detection was used to reflect the geometrical structural feature, not being the method of image segmentation. And the wavelet transform was introduced, which can avoid the difficulty of image understanding, efficiently show the distribution of image structure and improve the noise immunity in image retrieval.
本文重点研究以图像边缘特征和颜色特征为模型的检索算法,主要内容包括:首先介绍基于内容的图像检索技术,总结了颜色特征、纹理特征和形状特征的表达方式和提取方法,并对图像检索预处理的图像分割与像素分类作了研究,分析了现有检索算法的局限性与不足之处。
提出了两种综合利用图像边缘结构特征与颜色特征的图像检索算法。(1)基于Canny算子提取图像边缘与颜色目标定位的图像检索算法。(2)基于小波域提取图像边缘与颜色目标定位的图像检索算法。这两种算法分别利用Canny空间算子与小波变换域两种边缘提取方法对图像的形状特征进行描述,在待检索图像库中查找到与示例图像边缘相关度大的图像,构成一个新的检索图像库,然后再利用主颜色划分图像目标,利用有效的目标部分进行相似匹配,完成对示例图像的检索,仿真实验证明,这种结构的检索算法可以有效捕捉边缘信息,提高检索性能。
本文对边缘的提取不作为目标分割的手段,仅视为反映图像几何结构分布的低层次特征,而且引入小波域变换,避免了因图像理解的目标分割上的困难,有效地反映了图像的结构分布,提高了图像检索抗干扰性。
Content-Based Image Retrieval (CBIR) is an important research field in the development of image retrieval. It is an effective method of administering magnanimity digital image and video information. Its essence is to realize the image automatic search utilizing the special understanding way of the image. It can be used to realize one-to-many similar search.
This paper mainly studies retrieval algorithm based on image edge and color features. The main contents include:
First of all, the technology of CBIR was simply introduced. A great deal of methods of extraction and expression about the color, texture and shape features were summed up and summarized in detail. Image segmentation and pixels classifying of image retrieval pretreatment was studied. The limitation and deficiency of the existing retrieval algorithm was analyzed.
Second, two image retrieval algorithms using image edge structure and color features were put forward. (1) Image Retrieval Based on Canny Edge Detection and Goal Location. (2) Image Retrieval Based on the Wavelet Transform Edge Detection and Goal Location. The edge feature was described and extracted in the two kinds of domains: the spatial domain and the transform domain. The images that have the close correlation edge to the example image was searched to form a new multi-images database. Then it used the dominant color to divide image target, used effective target part for similarity matching and finishing the retrieval of the example images. Experiments show that this kind of arithmetic can well and efficiently detect the edge feature and improve the efficiency.
The edge detection was used to reflect the geometrical structural feature, not being the method of image segmentation. And the wavelet transform was introduced, which can avoid the difficulty of image understanding, efficiently show the distribution of image structure and improve the noise immunity in image retrieval.
引文
[1]Hua Xie,Antonio Ortega.Feature Representation and Compression for Content-based Retrieval.Proceeding of SPIE[J]-The International Society for Optical Engineering,2001,4310:111-122.
[2]P.W.Huang,S.K.Dai.Design of a Two-stage Content-based Image Retrieval System Using Texture Similarity[J].Information Processing and Management,2004,40(1):81-96.
[3]T.Kato,Databased architecture for content-based image retrieval.In Proceeding of SPIE[J]:Image Storage and Retrieval System,Vol.1662,112-123,1992.
[4]茹立云,彭潇,苏中.基于内容图像检索中的特征性能评价.计算机研究与发展[J].2003,40(11):1566-1570.
[5]胡晓峰,李国辉.多媒体系统[M].北京:人民邮电出版社.1997:4-9.
[6]夏德深,傅德胜.现代图像处理技术与应用[M].东南大学出版社.2001:33-35.
[7]John M Zachary Jr,Sitharama S Lyengar.Content-Based Image Retrieval System[J].IEEE Proceedings of the 1999 IEEE Symposium on Application-specific Systems and Software Engineering & Technology.1999.
[8]章毓晋.基于内容的视觉信息检索[M].科学出版社.2003(2):9-P13.
[9]陈琦.基于内容的图像检索中特征提取技术研究[D].硕士学位论文.浙江工业大学.2005.
[10]洪安祥.基于内容的图像检索若干论题研究[D].博士学位论文.浙江大学.2003.
[11]刘倩.基于内容的图像检索中的相关反馈技术[J].华东交通大学学报.2003(8).71-74.
[12]Swain M.J.Ballard D H.Color Indexing[J].International Journal of Computer Vision,1991,7(1):11-17.
[13]李弼程,彭天强,彭波.智能图像处理技术[M].北京电子工业出版社.2004:285-286.
[14]Kenneth R.Castleman:Digital Image Processing[M],电子工业出版社,2002:78-86
[15]李俊山,李旭辉.数字图像处理[M].清华大学出版社.2007.4.229-233.
[16]范立南,韩晓微,张文渊.图像处理与模式识别[M].科学出版社.2007.3.16-26.
[17]章毓晋:图像工程上册.图像处理[M]清华大学出版社 2003:50-65.
[18]章毓晋:基于内容的视觉信息检索[M]科学出版社 2003:15-18.
[19]施智平,胡宏,李清永.基于纹理普描述子的图像检索[J].软件学报.2005:16(78-80).
[20]董卫军,周明全,耿国华.基于纹理-形状特征的图像检索技术[J].计算机工程与 应用.2004,(24):9-11.
[21]洪子泉,杨静宇.统计模式识别中的特征提取[J].数据采集与处理.1991,6(2):38-43
[22]阮秋琦.数字图像处理学[M].北京.电子工业出版社.2001:21-35.
[23]J.Mao and A.K.Jain.Texture classification and segmentation using multiresolution simultaneous autoregressive models[J].1992.Pattern Recognition,Vol.25,No.2,pp:173-188.
[24]黄翔宇,章毓晋.基于压缩域的图像检索技术研究进展[J].中国图象图形学报.2003,8(5):499-508.
[25]王丽亚.图像的特征提取和分类[D].硕士学位论文.西安电子科技大学.2006.
[26]李凤云,郭全花.基于内容图像检索中的颜色分析和检索技术[J].河北建筑工程学院学报.2004年6月.Vol.22 No.2:6-9.
[27]韦娜.基于纹理特征的图像检索技术研究与现实[D].硕士学位论文.西北大学.2003.
[28]梅丽霞.基于区域的图像数据检索技术研究[D].硕士学位论文.2003.1.
[29]Swain M.J.,Ballard D H.Color indexing[J],Int.J.on Computer Vision 1991,7(1):11-32.
[30]Huang J.Image indexing using color correlograms[J].IEEE Computer Society Conference on Computer Visoion and Patterm Recognition.1996.8:792-796.
[31]Canny J.A Computation Approach to Edge Detection[J].IEEE-PAML,1986,8:679-698.
[32]Wynne Hsu,Chua Y.S.and Pung H.K.,An Integrated Color-Spatial Approach to Content-based Image Retrieval[J],Proc of the ACMMM Conf[C],199.305-313
[33]胡一君,邹北骥.基于分类子块的图像检索[J].计算机工程与科学.2006年第28卷第七期62-64.
[34]Albert Boggess & Francis J.Narcowich.A First Course in Wavelets with Fourier Analysi[D]s.电子工业出版社.2005:141-150.
[35]杨海军.图像检测中若干低级处理问题及在指纹识别中应用的研究[D].博士学位论文.西安交通大学.2001.
[36]蒋妙玲.小波域中图像增强的研究[D].硕士学位论文.大连海事大学.2005年.
[2]P.W.Huang,S.K.Dai.Design of a Two-stage Content-based Image Retrieval System Using Texture Similarity[J].Information Processing and Management,2004,40(1):81-96.
[3]T.Kato,Databased architecture for content-based image retrieval.In Proceeding of SPIE[J]:Image Storage and Retrieval System,Vol.1662,112-123,1992.
[4]茹立云,彭潇,苏中.基于内容图像检索中的特征性能评价.计算机研究与发展[J].2003,40(11):1566-1570.
[5]胡晓峰,李国辉.多媒体系统[M].北京:人民邮电出版社.1997:4-9.
[6]夏德深,傅德胜.现代图像处理技术与应用[M].东南大学出版社.2001:33-35.
[7]John M Zachary Jr,Sitharama S Lyengar.Content-Based Image Retrieval System[J].IEEE Proceedings of the 1999 IEEE Symposium on Application-specific Systems and Software Engineering & Technology.1999.
[8]章毓晋.基于内容的视觉信息检索[M].科学出版社.2003(2):9-P13.
[9]陈琦.基于内容的图像检索中特征提取技术研究[D].硕士学位论文.浙江工业大学.2005.
[10]洪安祥.基于内容的图像检索若干论题研究[D].博士学位论文.浙江大学.2003.
[11]刘倩.基于内容的图像检索中的相关反馈技术[J].华东交通大学学报.2003(8).71-74.
[12]Swain M.J.Ballard D H.Color Indexing[J].International Journal of Computer Vision,1991,7(1):11-17.
[13]李弼程,彭天强,彭波.智能图像处理技术[M].北京电子工业出版社.2004:285-286.
[14]Kenneth R.Castleman:Digital Image Processing[M],电子工业出版社,2002:78-86
[15]李俊山,李旭辉.数字图像处理[M].清华大学出版社.2007.4.229-233.
[16]范立南,韩晓微,张文渊.图像处理与模式识别[M].科学出版社.2007.3.16-26.
[17]章毓晋:图像工程上册.图像处理[M]清华大学出版社 2003:50-65.
[18]章毓晋:基于内容的视觉信息检索[M]科学出版社 2003:15-18.
[19]施智平,胡宏,李清永.基于纹理普描述子的图像检索[J].软件学报.2005:16(78-80).
[20]董卫军,周明全,耿国华.基于纹理-形状特征的图像检索技术[J].计算机工程与 应用.2004,(24):9-11.
[21]洪子泉,杨静宇.统计模式识别中的特征提取[J].数据采集与处理.1991,6(2):38-43
[22]阮秋琦.数字图像处理学[M].北京.电子工业出版社.2001:21-35.
[23]J.Mao and A.K.Jain.Texture classification and segmentation using multiresolution simultaneous autoregressive models[J].1992.Pattern Recognition,Vol.25,No.2,pp:173-188.
[24]黄翔宇,章毓晋.基于压缩域的图像检索技术研究进展[J].中国图象图形学报.2003,8(5):499-508.
[25]王丽亚.图像的特征提取和分类[D].硕士学位论文.西安电子科技大学.2006.
[26]李凤云,郭全花.基于内容图像检索中的颜色分析和检索技术[J].河北建筑工程学院学报.2004年6月.Vol.22 No.2:6-9.
[27]韦娜.基于纹理特征的图像检索技术研究与现实[D].硕士学位论文.西北大学.2003.
[28]梅丽霞.基于区域的图像数据检索技术研究[D].硕士学位论文.2003.1.
[29]Swain M.J.,Ballard D H.Color indexing[J],Int.J.on Computer Vision 1991,7(1):11-32.
[30]Huang J.Image indexing using color correlograms[J].IEEE Computer Society Conference on Computer Visoion and Patterm Recognition.1996.8:792-796.
[31]Canny J.A Computation Approach to Edge Detection[J].IEEE-PAML,1986,8:679-698.
[32]Wynne Hsu,Chua Y.S.and Pung H.K.,An Integrated Color-Spatial Approach to Content-based Image Retrieval[J],Proc of the ACMMM Conf[C],199.305-313
[33]胡一君,邹北骥.基于分类子块的图像检索[J].计算机工程与科学.2006年第28卷第七期62-64.
[34]Albert Boggess & Francis J.Narcowich.A First Course in Wavelets with Fourier Analysi[D]s.电子工业出版社.2005:141-150.
[35]杨海军.图像检测中若干低级处理问题及在指纹识别中应用的研究[D].博士学位论文.西安交通大学.2001.
[36]蒋妙玲.小波域中图像增强的研究[D].硕士学位论文.大连海事大学.2005年.