自主移动机器人环境图像识别方法研究
|
摘要
伴随着计算机技术的不断开拓和对人类感知机理研究的不断发展,智能移动机器人的先进控制成为了学术界关注的热点。赋予机器人“看”的能力是机器人视觉研究的主要问题之一。机器人利用视觉实现对环境的识别是一个复杂的过程,因此,研究识别率高、抗噪能力强、鲁棒性好的图像识别方法对移动机器人研究领域具有重要意义和理论价值。
本文首先对机器人采集环境图像的预处理方法进行了研究：
(1)研究了对原始图像的增强方法。采用分段线性函数和直方图处理对光线暗的图像增强对比度；然后对含噪声的模糊图像进行去噪研究,对比了常用的几种去噪方法,并采用了一种改进的阈值折中小波去噪的方法进行降噪。实验证明,该方法能够得到较理想的降噪效果。
(2)采用了不同的模板对图像进行边缘提取以及二值化处理,并进行了实验比较及分析,为后续图像识别做好了充分准备。
其次,对环境图像识别、分类的方法进行了研究：
(1)采用SIFT (Scale-invariant feature transform)特征提取的方法,建立图像金字塔,提取图像的边缘特征,然后采用特征点之间的最近欧氏距离进行识别。而后在此基础上,本文从实时性方面进行了改进：①在特征点进行欧氏距离的匹配的时候加入一个阈值,将那些不匹配的点可以直接滤除,不必进行复杂的128次差平方运算；②在建立图像金字塔的时候没必要一直进行到最后一组足够小为止,本文通过实验得到第4组的时候特征点的总数就不再增加了,不需要一直计算下去,所以节省了运行的时间。实验证明了这两种方法的有效性。
(2)本文还研究了基于Gabor滤波器和不变矩的神经网络的图像分类方法,研究了Gabor滤波器以及不变矩提取图像特征的原理,使用Gabor滤波器组提取32个局部特征,不变矩提取7个全局特征,将特征值通过BP神经网络学习,最终实现对环境图像的分类。仿真实验验证了该算法提高了识别精度的同时也提高了识别的鲁棒性,满足智能移动机器人对环境识别的要求。
With the development of computer technologies and human sensation researches,the advanced control of intelligent robot has become a hot spot in academic.Giving the robot the function of look is one of the main questions of robot vision research. And the robot use vision to recognition the environment is a complex process.Therefore, the research of real-time vision recognition system with a high recognition rate, a strong anti-noiseability and highly robust performance has a vital significance and the theoretical value.
First of all in this dissertation,the environment image'pretreatment methods are studied:
(1) The enhancement methods of original environment images are studied.Though using piecewise-linear function and histogram processing,the light-dark images'contrast are enchanced.Then, for the fuzzy images with noise,comparing several general used de-noising method,an improved wavelet mid-threshold denoising method is proposed.The experiment result shows that this method can achieve superior de-noising performance.
(2) Image edge extraction based on a variety of templates and binarization processing methods are studied,and the experiment result is compared and analyzed. This process is well prepared for follow-up image recognition.
Secondly,the environment images recognition and classification methods are studied:
(1) Using SIFT(Scale-invariant feature transform)feature extraction methods, building the image pyramid, extracting image's edge features and then the recognition between the two images is though their features' Euclidean distance.In addition,on the basis of SIFT,the dissertation has two improvemwnt in terms of real time:first,a new threshold is added to the process of the match between the two features,in order to exclude directly the features which are not matched,so that the complex 128 deviation square operation is no need to be calculated.Second,it is not need to continue until the last enough small group when the image pyramid is established,and though the experiment it is found that the total number of features will no longer be increased over the first 4 group.Therefore,this method saves running time.Simulation experimental results show the two methods are effectiveness.
(2) Neural network classification method based on Gabor filter and moment invariant is studied in this dissertation.The principle of image feature extraction based on Gabor filter as well as moment invariant is described.32 local features are extracted though Gabor filter and 7 global features are extracted though moment invariant.Then,the 39 features value is learned by neural network and adjustable parameters of neural network are trained to optimize.Simulation results show that the algorithm improves the recognition accuracy at the same time also cuts down the recognition time.This algorithm is able to meet the requirement to recognize the environment for intelligent mobile.
本文首先对机器人采集环境图像的预处理方法进行了研究：
(1)研究了对原始图像的增强方法。采用分段线性函数和直方图处理对光线暗的图像增强对比度；然后对含噪声的模糊图像进行去噪研究,对比了常用的几种去噪方法,并采用了一种改进的阈值折中小波去噪的方法进行降噪。实验证明,该方法能够得到较理想的降噪效果。
(2)采用了不同的模板对图像进行边缘提取以及二值化处理,并进行了实验比较及分析,为后续图像识别做好了充分准备。
其次,对环境图像识别、分类的方法进行了研究：
(1)采用SIFT (Scale-invariant feature transform)特征提取的方法,建立图像金字塔,提取图像的边缘特征,然后采用特征点之间的最近欧氏距离进行识别。而后在此基础上,本文从实时性方面进行了改进：①在特征点进行欧氏距离的匹配的时候加入一个阈值,将那些不匹配的点可以直接滤除,不必进行复杂的128次差平方运算；②在建立图像金字塔的时候没必要一直进行到最后一组足够小为止,本文通过实验得到第4组的时候特征点的总数就不再增加了,不需要一直计算下去,所以节省了运行的时间。实验证明了这两种方法的有效性。
(2)本文还研究了基于Gabor滤波器和不变矩的神经网络的图像分类方法,研究了Gabor滤波器以及不变矩提取图像特征的原理,使用Gabor滤波器组提取32个局部特征,不变矩提取7个全局特征,将特征值通过BP神经网络学习,最终实现对环境图像的分类。仿真实验验证了该算法提高了识别精度的同时也提高了识别的鲁棒性,满足智能移动机器人对环境识别的要求。
With the development of computer technologies and human sensation researches,the advanced control of intelligent robot has become a hot spot in academic.Giving the robot the function of look is one of the main questions of robot vision research. And the robot use vision to recognition the environment is a complex process.Therefore, the research of real-time vision recognition system with a high recognition rate, a strong anti-noiseability and highly robust performance has a vital significance and the theoretical value.
First of all in this dissertation,the environment image'pretreatment methods are studied:
(1) The enhancement methods of original environment images are studied.Though using piecewise-linear function and histogram processing,the light-dark images'contrast are enchanced.Then, for the fuzzy images with noise,comparing several general used de-noising method,an improved wavelet mid-threshold denoising method is proposed.The experiment result shows that this method can achieve superior de-noising performance.
(2) Image edge extraction based on a variety of templates and binarization processing methods are studied,and the experiment result is compared and analyzed. This process is well prepared for follow-up image recognition.
Secondly,the environment images recognition and classification methods are studied:
(1) Using SIFT(Scale-invariant feature transform)feature extraction methods, building the image pyramid, extracting image's edge features and then the recognition between the two images is though their features' Euclidean distance.In addition,on the basis of SIFT,the dissertation has two improvemwnt in terms of real time:first,a new threshold is added to the process of the match between the two features,in order to exclude directly the features which are not matched,so that the complex 128 deviation square operation is no need to be calculated.Second,it is not need to continue until the last enough small group when the image pyramid is established,and though the experiment it is found that the total number of features will no longer be increased over the first 4 group.Therefore,this method saves running time.Simulation experimental results show the two methods are effectiveness.
(2) Neural network classification method based on Gabor filter and moment invariant is studied in this dissertation.The principle of image feature extraction based on Gabor filter as well as moment invariant is described.32 local features are extracted though Gabor filter and 7 global features are extracted though moment invariant.Then,the 39 features value is learned by neural network and adjustable parameters of neural network are trained to optimize.Simulation results show that the algorithm improves the recognition accuracy at the same time also cuts down the recognition time.This algorithm is able to meet the requirement to recognize the environment for intelligent mobile.
引文
[1]张毅,罗元,郑太雄.移动机器人及其应用[M].北京：电子工业出版社,2007.
[2]张广军.机器视觉[M].北京：科学出版社,2005.
[3]陈亮.Gabor小波特征提取技术及其在目标识别中的应用研究[D].南京：南京理工大学,2009.
[4]关晓胜.机器视觉及其应用发展[J].新型仪器与智能仪表,2005,8(46)：120-122.
[5]刘进.不变量特征的构造及在目标识别中的应用[D].武汉：华中科技大学,2004.
[6]桑克(Sonka,M).图像处理、分析和机器视觉[M].北京：人民邮电出版社,2009.
[7]张会章,张立霞.图像理解[J].计算机工程,2003,7(13)：88-90.
[8]蒋先刚.数字图像模式识别工程软件设计[M].北京：中国水利水电出版社,2008.
[9]冈萨雷斯.数字图像处理[M].北京：电子工业出版社,2006.
[10]傅京孙,程民德译.模式识别应用[M].北京：北京大学出版社,1990.
[11]陈强.基于组合矩和神经网络的图像识别技术研究[D].南京：南京理工大学,2007.
[12]孙即祥.现代模式识别[M].长沙：国防科技大学出版社,2002.
[13]许延伟,刘希玉.神经网络用于模式识别的几种常见方法及比较[J].信息技术与信息化,2005,(4)：120-123.
[14]杨治明.人工神经网络及其在图像识别中的应用研究[D].重庆：重庆大学,2003.4.
[15]赵鹏.汽车铝轮毂X射线检测的图像处理技术研究[D].山西：中北大学,2007.
[16]Hearn D,Baker MP.Computer graphics[M].New Jersey:Prentice Hall Inc,1997.
[17]P K Sahoo,S Solana,AWong et al.A survey of thresholding techniques [J].Computer Vision,Graphics and Image Processing,1998, (41):233-260.
[18]陈杨,陈荣娟,郭颖辉等.Matlab6.X图形编程与图像处理[M].西安：西安电子科技大学出版社,2002.
[19]阿查里雅.数字图像处理原理与应用[M].北京：清华大学出版社,2007.
[20]贺兴华,周媛媛.Matlab7.X图像处理[M].北京：人民邮电出版社,2006.
[21]艾海舟,武勃.图像处理、分析与机器视觉[M].北京：人民邮电出版社,2003.
[22]王晅.图像旋转与尺度变换不变性识别方法研究[D].西安：西安电子科技大学,2008.
[23]应义赋,景寒松,马俊福等.黄花梨果形的机器视觉视觉识别方法研究[J].农业工程学报,1999,15(1)：192-196.
[24]孙即祥.图像分析[M].北京：科学出版社,2005.
[25]Wang X.Nonlinear multi-wavelet transform based soft thresholding[A].In Conf IEEE A PCCAS[C]. Tian jin:2000:775-778.
[26]杨淑莹.VC++图像处理程序设计[M].北京：电子工业出版社,1993.
[27]T Kurita,N Otsu,N Abdelmalek.Maximum likelihood threshold based on population mixture models[J].Pattern Recognition,1992,25 (10):1231-1240.
[28]傅景广,徐刚,王裕国.基于模式化方法的二值图像降噪[J].计算机工程与应用,2003,15(33)：220-222.
[29]J D Yang S Chen,W H Hsu.Adaptive threshold algorithm and its hardware implementation[J]. Patte-rn Recognition Letters,1994, (15):141-150.
[30]汪成为.灵境技术的理论、实践及应用[M].南宁：广西科学技术出版社,1996.
[31]钟力,胡晓峰.重叠图像拼接法[J].中国图像图形学报,1998,(3)：365-369.
[32]Kuglin C D,Hines D.C.The Phase Correlation Image alignment method[A].IEEE Conference on Cybe-rnetics and Society [C].New York:1975:163-165.
[33]David G.Lowe.Object recognition from local sale-invariant features[A].Internitional Conference on Computer Vision[C].New York:1999:1150-1157.
[34]David G.Lowe.Distinctive Image Features from Sale-Invariant keypoints[J].Internitional Journal of Computer Vision,2004, (3):91-110.
[35]W.F.Bischof and T.Caclli.Parsing scale-space and spatial stability analysis[J].Computer Vision, Graphics and Image Processing,1988, (42):192-205.
[36]Mikplajczyk K,Schmid C.A performance evaluation of local descriptor[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2005,27(10):1615-1630.
[37]Koenderink.The structure of images[J].Biological Cybernetics,1984, (55):363-396.
[38]Lindeberg.Scale-space for discrete signals[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1980, (20):7-18.
[39]张春美,龚志辉,孙雷.改进SIFT特征在图像匹配中的应用[J].计算机工程与应用,2008,44(2)：95-97.
[40]季方.改进山歌交通地图道路识别过程的研究[D].北京：中国科技大学,2005.
[41]田阿灵.基于SIFT的红外与可见光图像配准方法研究[D].北京：华北电力大学,2008.
[42]R Bellman,R Kalaba.Adaptive Control Process[J].IRE Transactions on Automatic Control,1959, 4 (2):1-9.
[43]林传力.基于内容的图像检索和SIFT算法的应用[D].上海：上海交通大学,2008.
[44]骞森,朱剑英.基于改进的SIFT特征的图像双向匹配算法[J].机械科学与技术,2007,26(9)：1179-1182.
[45]董海鹰.智能控制理论与应用[M].北京：中国铁道出版社,2006.
[46]Richard O.Duda,Peter E.Hart,David G.Stork.Pattern Classification[M].John Wiley & Sons,Inc,2001.
[47]郭立,陆大虎,朱俊株.基于Gabor多通道滤波和Hopfield神经网络的文理图像分割[J].计算机工程与应用,2000,(6)：39-41.
[48]Miaoli Wen,Yan Liang,Quan Pan,Hongcai Zhang.A Gabor filter based fingerprint enchancement algorithm in wavelet domain[J].IEEE.international symposium on communicationa and information technology,2005, (2):1468-1471.
[49]Areekul V.,Watchareeruetai U.,Suppasriwasuseth K.,Tantaratana S.Separable Gabor filter realization for fast fingerprint enchance[A].IEEE International Conference on Image Processing[C].New York: 2005, (3):253-256.
[50]Khmanee C.,Nguyen D.On the design of 2D Gabor filter of fingerprint images[A].Consumers Com-munication and Networking Conference[C].New York:2004, (1):430-435.
[51]Manjunath B,Ma W Y.Textures features for browsing and retrieval of image data[J].IEEE Transactio- ns on Pattern Analysis and Machine Intelligence,1996,18 (8):336-345.
[52]M K Hu.Visual Pattern Recognition by Moment Invaiant[J].IEEE Transactions on Information Theory, 1962,8(2):179-187.
[53]Mukundan R,Ramakrishnan K R.Fast computation of Lgendre and Zernike moments[J].Pattern Recog-nition,1995,28 (9):1433-1442.
[54]沈定刚.用于图像目标识别的神经网络方法[J].系统工程与电子技术,1994,13(25)：69-71.
[55]陈明.神经网络模型[M].大连：大连理工出版社,1995.
[56]楼顺天,施阳.基于MATLAB的系统分析与设计—神经网络[M].西安：西安电子科技大学出版社,1998.
[57]周开利,康耀红.神经网络模型及其MATLAB仿真程序设计[M].北京：清华大学出版社,2005.
[58]赵廷钊.基于细胞神经网络的煤仓图像处理方法的研究[J].中国煤炭,2010,(01)：73-78.
[2]张广军.机器视觉[M].北京：科学出版社,2005.
[3]陈亮.Gabor小波特征提取技术及其在目标识别中的应用研究[D].南京：南京理工大学,2009.
[4]关晓胜.机器视觉及其应用发展[J].新型仪器与智能仪表,2005,8(46)：120-122.
[5]刘进.不变量特征的构造及在目标识别中的应用[D].武汉：华中科技大学,2004.
[6]桑克(Sonka,M).图像处理、分析和机器视觉[M].北京：人民邮电出版社,2009.
[7]张会章,张立霞.图像理解[J].计算机工程,2003,7(13)：88-90.
[8]蒋先刚.数字图像模式识别工程软件设计[M].北京：中国水利水电出版社,2008.
[9]冈萨雷斯.数字图像处理[M].北京：电子工业出版社,2006.
[10]傅京孙,程民德译.模式识别应用[M].北京：北京大学出版社,1990.
[11]陈强.基于组合矩和神经网络的图像识别技术研究[D].南京：南京理工大学,2007.
[12]孙即祥.现代模式识别[M].长沙：国防科技大学出版社,2002.
[13]许延伟,刘希玉.神经网络用于模式识别的几种常见方法及比较[J].信息技术与信息化,2005,(4)：120-123.
[14]杨治明.人工神经网络及其在图像识别中的应用研究[D].重庆：重庆大学,2003.4.
[15]赵鹏.汽车铝轮毂X射线检测的图像处理技术研究[D].山西：中北大学,2007.
[16]Hearn D,Baker MP.Computer graphics[M].New Jersey:Prentice Hall Inc,1997.
[17]P K Sahoo,S Solana,AWong et al.A survey of thresholding techniques [J].Computer Vision,Graphics and Image Processing,1998, (41):233-260.
[18]陈杨,陈荣娟,郭颖辉等.Matlab6.X图形编程与图像处理[M].西安：西安电子科技大学出版社,2002.
[19]阿查里雅.数字图像处理原理与应用[M].北京：清华大学出版社,2007.
[20]贺兴华,周媛媛.Matlab7.X图像处理[M].北京：人民邮电出版社,2006.
[21]艾海舟,武勃.图像处理、分析与机器视觉[M].北京：人民邮电出版社,2003.
[22]王晅.图像旋转与尺度变换不变性识别方法研究[D].西安：西安电子科技大学,2008.
[23]应义赋,景寒松,马俊福等.黄花梨果形的机器视觉视觉识别方法研究[J].农业工程学报,1999,15(1)：192-196.
[24]孙即祥.图像分析[M].北京：科学出版社,2005.
[25]Wang X.Nonlinear multi-wavelet transform based soft thresholding[A].In Conf IEEE A PCCAS[C]. Tian jin:2000:775-778.
[26]杨淑莹.VC++图像处理程序设计[M].北京：电子工业出版社,1993.
[27]T Kurita,N Otsu,N Abdelmalek.Maximum likelihood threshold based on population mixture models[J].Pattern Recognition,1992,25 (10):1231-1240.
[28]傅景广,徐刚,王裕国.基于模式化方法的二值图像降噪[J].计算机工程与应用,2003,15(33)：220-222.
[29]J D Yang S Chen,W H Hsu.Adaptive threshold algorithm and its hardware implementation[J]. Patte-rn Recognition Letters,1994, (15):141-150.
[30]汪成为.灵境技术的理论、实践及应用[M].南宁：广西科学技术出版社,1996.
[31]钟力,胡晓峰.重叠图像拼接法[J].中国图像图形学报,1998,(3)：365-369.
[32]Kuglin C D,Hines D.C.The Phase Correlation Image alignment method[A].IEEE Conference on Cybe-rnetics and Society [C].New York:1975:163-165.
[33]David G.Lowe.Object recognition from local sale-invariant features[A].Internitional Conference on Computer Vision[C].New York:1999:1150-1157.
[34]David G.Lowe.Distinctive Image Features from Sale-Invariant keypoints[J].Internitional Journal of Computer Vision,2004, (3):91-110.
[35]W.F.Bischof and T.Caclli.Parsing scale-space and spatial stability analysis[J].Computer Vision, Graphics and Image Processing,1988, (42):192-205.
[36]Mikplajczyk K,Schmid C.A performance evaluation of local descriptor[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2005,27(10):1615-1630.
[37]Koenderink.The structure of images[J].Biological Cybernetics,1984, (55):363-396.
[38]Lindeberg.Scale-space for discrete signals[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1980, (20):7-18.
[39]张春美,龚志辉,孙雷.改进SIFT特征在图像匹配中的应用[J].计算机工程与应用,2008,44(2)：95-97.
[40]季方.改进山歌交通地图道路识别过程的研究[D].北京：中国科技大学,2005.
[41]田阿灵.基于SIFT的红外与可见光图像配准方法研究[D].北京：华北电力大学,2008.
[42]R Bellman,R Kalaba.Adaptive Control Process[J].IRE Transactions on Automatic Control,1959, 4 (2):1-9.
[43]林传力.基于内容的图像检索和SIFT算法的应用[D].上海：上海交通大学,2008.
[44]骞森,朱剑英.基于改进的SIFT特征的图像双向匹配算法[J].机械科学与技术,2007,26(9)：1179-1182.
[45]董海鹰.智能控制理论与应用[M].北京：中国铁道出版社,2006.
[46]Richard O.Duda,Peter E.Hart,David G.Stork.Pattern Classification[M].John Wiley & Sons,Inc,2001.
[47]郭立,陆大虎,朱俊株.基于Gabor多通道滤波和Hopfield神经网络的文理图像分割[J].计算机工程与应用,2000,(6)：39-41.
[48]Miaoli Wen,Yan Liang,Quan Pan,Hongcai Zhang.A Gabor filter based fingerprint enchancement algorithm in wavelet domain[J].IEEE.international symposium on communicationa and information technology,2005, (2):1468-1471.
[49]Areekul V.,Watchareeruetai U.,Suppasriwasuseth K.,Tantaratana S.Separable Gabor filter realization for fast fingerprint enchance[A].IEEE International Conference on Image Processing[C].New York: 2005, (3):253-256.
[50]Khmanee C.,Nguyen D.On the design of 2D Gabor filter of fingerprint images[A].Consumers Com-munication and Networking Conference[C].New York:2004, (1):430-435.
[51]Manjunath B,Ma W Y.Textures features for browsing and retrieval of image data[J].IEEE Transactio- ns on Pattern Analysis and Machine Intelligence,1996,18 (8):336-345.
[52]M K Hu.Visual Pattern Recognition by Moment Invaiant[J].IEEE Transactions on Information Theory, 1962,8(2):179-187.
[53]Mukundan R,Ramakrishnan K R.Fast computation of Lgendre and Zernike moments[J].Pattern Recog-nition,1995,28 (9):1433-1442.
[54]沈定刚.用于图像目标识别的神经网络方法[J].系统工程与电子技术,1994,13(25)：69-71.
[55]陈明.神经网络模型[M].大连：大连理工出版社,1995.
[56]楼顺天,施阳.基于MATLAB的系统分析与设计—神经网络[M].西安：西安电子科技大学出版社,1998.
[57]周开利,康耀红.神经网络模型及其MATLAB仿真程序设计[M].北京：清华大学出版社,2005.
[58]赵廷钊.基于细胞神经网络的煤仓图像处理方法的研究[J].中国煤炭,2010,(01)：73-78.