航空发动机滑油滤磨屑图像检测与识别技术研究
|
摘要
随着机械工业和航空技术的不断发展,航空发动机的结构日益复杂,其润滑系统中各摩擦副零组件更趋于高载荷、高温、高速及轻质量,更加容易发生各种磨损故障,从而严重影响发动机的安全性和可靠性。润滑系统滑油滤上的金属屑,包含了大量反映发动机工作状态是否正常的有用信息,滑油滤监控已成为国内外新一代发动机监控的发展趋势。研究发动机滑油滤磨屑监控方法,建立定量化、标准化的便于操作的工艺方法和监控标准,提高发动机润滑系统故障预报成功率,对于解决发动机安全使用的难题及完成战训飞行任务具有重要意义。鉴于此,本文进行了基于滑油滤图像分析的航空发动机磨损状态自动检测系统研究。
第一、阐述了典型航空发动机润滑系统的组成及其工作原理;分析了发动机磨损形成机理;并介绍了航空发动机滑油滤磨屑检测原理及方法。
第二、本文研究了滑油滤磨屑磨损特征的提取技术。针对滑油滤磨屑图像,首先运用图像自动阈值分割技术提取磨屑面积特征量;其次运用数学形态学方法提取磨屑堆积面积;最后,运用小波边缘检测技术提取磨屑周长特征量。
第三、运用多元回归方法进行磨损烈度识别。该方法利用磨屑图像提取出来的磨损特征量及相关图像的磨损烈度值进行多元回归分析,并针对回归结果进行假设性检验。最后,用算例验证了该方法的有效性。
第四、提出了一种基于野点检测的磨屑图像识别方法。该方法通过对正常磨屑案例图像特征量求取正常域边界,建立野点检测模型。有效地解决了滑油滤磨屑图像正常样本获取容易,异常样本获取困难的问题。最后,通过算例验证了该方法的有效性。
第五、利用Microsoft Visual C++6.0和Microsoft Access 2003数据库开发了发动机滑油滤监控系统EOFMS(Engine Oil Filter Monitor System)。构建了系统的整体架构,实现了基于多元回归的磨屑图像诊断功能和基于野点检测的磨屑图像识别功能等。
With the development of machinery industry and aviation technology, the structure of aero-engine is getting more and more complicated, and parts with frication pairs of lubrication system tend to be higher load, temperature, speed and lighter mass. So various kinds of wear fault are happened easily, and seriously affect the safety and reliability of aero-engine. The debris on oil filter contains lots of useful information which can reflect the aero-engine's working state. The oil filter monitoring is becoming a trend of new engine monitoring at home and abroad. Studying on oil filter monitoring methods, setting up convenient operation techniques and monitoring standards, and improving the success rate of fault prediction for lubrication system, have great significance in solving the safety use problems of aero-engine and completing the flight mission. Therefore, in this paper, an automatic detection system of aero-engine wear condition based on debris image analysis is studied.
Firstly, the composition and working principle of typical aero-engine lubrication system are described, wear mechanism of aero-engine is analyzed, and the Principles and Methods of oil filter debris detection are introduced.
Secondly, in this paper, the feature extraction techniques of oil filter debris are studied. For oil filter debris images, we first obtain debris area characteristic quantity using automatic threshold segmentation. Then extract the debris accumulation based on mathematical morphology method. Finally, we get the debris circumference characteristic quantity by applying wavelet edge detection. Thirdly, multiple regression method is used to analyze wear intensity. In this method, multiple regression analysis and hypothesis testing of regression results are carried out, based on the wear characteristic quantity and wear intensity of debris images. Finally, the valid of the method is proved by examples.
Fourthly, a method for debris image recognition based on novelty detection is put forward. This method can obtain the normal domain boundary from the characteristic quantity of normal case images, and establish a dynamic model of novelty detection. The method has effectively solved the problem of easily possessing normal samples and hardly obtaining abnormal samples. Finally, the valid of the method is proved by examples.
Fifthly, an Engine Oil Filter Monitor System is developed by using the development tools of Microsoft Visual C++6.0 language and Microsoft Access 2003 database. The whole structure of the system is constructed, and the functions of debris image diagnosis based on multiple regression and novelty detection etc are realized.
第一、阐述了典型航空发动机润滑系统的组成及其工作原理;分析了发动机磨损形成机理;并介绍了航空发动机滑油滤磨屑检测原理及方法。
第二、本文研究了滑油滤磨屑磨损特征的提取技术。针对滑油滤磨屑图像,首先运用图像自动阈值分割技术提取磨屑面积特征量;其次运用数学形态学方法提取磨屑堆积面积;最后,运用小波边缘检测技术提取磨屑周长特征量。
第三、运用多元回归方法进行磨损烈度识别。该方法利用磨屑图像提取出来的磨损特征量及相关图像的磨损烈度值进行多元回归分析,并针对回归结果进行假设性检验。最后,用算例验证了该方法的有效性。
第四、提出了一种基于野点检测的磨屑图像识别方法。该方法通过对正常磨屑案例图像特征量求取正常域边界,建立野点检测模型。有效地解决了滑油滤磨屑图像正常样本获取容易,异常样本获取困难的问题。最后,通过算例验证了该方法的有效性。
第五、利用Microsoft Visual C++6.0和Microsoft Access 2003数据库开发了发动机滑油滤监控系统EOFMS(Engine Oil Filter Monitor System)。构建了系统的整体架构,实现了基于多元回归的磨屑图像诊断功能和基于野点检测的磨屑图像识别功能等。
With the development of machinery industry and aviation technology, the structure of aero-engine is getting more and more complicated, and parts with frication pairs of lubrication system tend to be higher load, temperature, speed and lighter mass. So various kinds of wear fault are happened easily, and seriously affect the safety and reliability of aero-engine. The debris on oil filter contains lots of useful information which can reflect the aero-engine's working state. The oil filter monitoring is becoming a trend of new engine monitoring at home and abroad. Studying on oil filter monitoring methods, setting up convenient operation techniques and monitoring standards, and improving the success rate of fault prediction for lubrication system, have great significance in solving the safety use problems of aero-engine and completing the flight mission. Therefore, in this paper, an automatic detection system of aero-engine wear condition based on debris image analysis is studied.
Firstly, the composition and working principle of typical aero-engine lubrication system are described, wear mechanism of aero-engine is analyzed, and the Principles and Methods of oil filter debris detection are introduced.
Secondly, in this paper, the feature extraction techniques of oil filter debris are studied. For oil filter debris images, we first obtain debris area characteristic quantity using automatic threshold segmentation. Then extract the debris accumulation based on mathematical morphology method. Finally, we get the debris circumference characteristic quantity by applying wavelet edge detection. Thirdly, multiple regression method is used to analyze wear intensity. In this method, multiple regression analysis and hypothesis testing of regression results are carried out, based on the wear characteristic quantity and wear intensity of debris images. Finally, the valid of the method is proved by examples.
Fourthly, a method for debris image recognition based on novelty detection is put forward. This method can obtain the normal domain boundary from the characteristic quantity of normal case images, and establish a dynamic model of novelty detection. The method has effectively solved the problem of easily possessing normal samples and hardly obtaining abnormal samples. Finally, the valid of the method is proved by examples.
Fifthly, an Engine Oil Filter Monitor System is developed by using the development tools of Microsoft Visual C++6.0 language and Microsoft Access 2003 database. The whole structure of the system is constructed, and the functions of debris image diagnosis based on multiple regression and novelty detection etc are realized.
引文
[1]左洪福.发动机磨损状态检测与故障诊断技术[M].北京:航空工业出版社, 1996.
[2]陈果.航空器检测与诊断技术导论[M].中国民航出版社, 2007.
[3]杨春生,孟昭荣编译.世界航空安全与事故分析[M].北京:中国民航出版社, 1997.
[4]吴振锋.基于磨粒分析和信息融合的发动机磨损故障诊断技术研究[D].南京航空航天大学, 2001.
[5]范作民,孙春林,白杰.航空发动机故障诊断导论[M].科学出版社, 2004.
[6]陶春虎,习年生,钟培道.航空装备失效典型案例分析[M].北京:国防工业出版社, 1998.
[7] D. P. Anderson著.金元生,杨其明译.磨粒图谱[M].北京:机械工业出版社, 1987.
[8] Tzafestas S. G, Konstantinidis N. I. ENGEXP-an Integrated Environment for The Development and Application of Expert Systems in Equipment and Engine Fault Diagnosis and Repair[J]. Advances in Engineering Software and workstations, 1992, 14(1):3-14.
[9] Raadnui S, Roylance B. J. The classification of wear particle shape[J]. Lubrication Engineering, 1995, 51(5):432-437.
[10] K. K. Yeung, et al. Development of computer-aided image analysis for filter debris analysis[J]. Lubrication Engineering, 1994, 50(4):293-299.
[11] F. T. Barwell. The contribution of particle analysis to the study of wear of metals[J]. Wear, 1983,90:167-181.
[12] A. D. H. Thomas, et al. Computer image analysis for identification of wear Particles[J].Wear, 1991, 142(1):213-226.
[13] Tomasz Pancewicz, Irena Mruk. Holographic contouring for determination of three- dimensional description of surface roughness[J]. Wear, 1996, 199(1):127-131.
[14] T. B. Kirk, et al. Fractal Parameters and computer image analysis applied to wear particles isolated by ferrography[J]. Wear, 1991, 145(2):347-365.
[15] Kun Xu, A. R. Luxmoore. An integrated system for automatic wear particle analysis[J]. Wear, 1997, 208(1):184-193.
[16] Podsiadlo P, Stachowiak G. W. Scale-invariant analysis of wear particle surface morphology[J]. Wear, 2000, 24(2):160-201.
[17] Laghari M. S, Memon Q. A. Knowledge based wear particle analysis[J]. International Journal of Information Technology, 2004, 1(3):91-95.
[18]谢友柏.电磁铁谱仪与铁谱技术的研究[R].西安:西安交通大学科技报告, 1983.
[19]杨志伊.铁谱仪的新发展-KTP旋转式铁谱仪[J].设备管理与维修,1990, 7:22~24.
[20]萧汉梁著.铁谱技术及其在机械监测诊断中的应用[M].北京:人民交通出版社, 1993.
[21]陈国安,葛世荣.基于分形理论的磨合磨损预测模型[J].机械工程学报, 2000, 36(2):29-33.
[22]李忠,曾昭翔.基于BP神经网络的磨损微粒智能识别[J].北方交通大学学报, 1998, 22(1):86-91.
[23]涂群章,左洪福,李艳军.发动机磨损颗粒图像在线检测技术[J].内燃机工程, 2003, 24(5):70-73.
[24]王伟华,殷勇辉.基于径向基函数神经网络的磨粒识别系统[J].摩擦学学报, 2003, 123(4):340-343.
[25]姜旭峰,彭秀华.基于灰色关联分析的磨粒识别[J].润滑与密封, 2005, 171(5):112-116.
[26]高经纬.遗传算法和神经网络在铁谱图像识别中的应用[J].润滑与密封,2004,161(1):16-18.
[27]钟新辉.粗糙集和证据理论在磨粒识别中的应用[J].润滑与密封, 2005, 5:143-145.
[28]康剑莉,芦亚萍,周银生.改进的BP算法在磨粒识别中的应用[J].润滑与密封, 2005, 2:149-154.
[29]李华强,费逸伟.基于Matlab聚类分析的磨粒分类识别研究[J].润滑与密封,2005,3:84-85.
[30]顾大强.基于支持向量机的铁谱磨粒模式识别[J].中国机械工程, 2006, 17(13):1391-1394.
[31]刘国光.基于改进支持向量机的磨粒铁谱识别的研究[J].润滑与密封, 2005, 3:94-98.
[32]左洪福,李立.发动机磨屑尺寸的分布模型及其应用[J].南京航空航天大学学报, 1995, 27(6):784-791.
[33] H. E. Powrie, C. E. Fisher. Engine health monitoring: Towards total prognostics[J]. Aerospace Conference, 1999, 3:11-20.
[34] Powrie H. E. G, McNicholas K. Gas path condition monitoring during Accelerated Mission Testing of a demonstrator engine[C]. The 33rd AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit Seattle. WA, USA: AIAA, 1997: 2904-2914.
[35] Nurse J, Petch C, Fisher C. E. Engine gas path integrity monitoring[C]. Aerotech 94, 1994(7):C470/6/052.
[36]陈果.航空发动机磨损故障的智能融合诊断[J].中国机械工程, 2005, 16(4):299-306.
[37] Tauber T, Johnsen D. Experience with the electronic oil debris monitoring system of the General Electric GE90 gas turbine engine on the Boeing 777 aircraft[J], JOAP International condition monitoring conference, Mobile, Alabama, 2002.
[38] Saba C. S, Wolf J. D, Dell A. J. XRF, a choice method for metal analysis[J], JOAP Internationalcondition monitoring conference, Mobile, Alabama, 2002.
[39] Lee R. J, Toms A. M. An SEM approach to wear debris analysis[J], JOAP International condition monitoring conference, Mobile, Alabama, 2002.
[40] Humphrey G. R, Gorzen D. Prototype instruments built for PRAM[J], JOAP International condition monitoring conference, Mobile, Alabama, 2002.
[41]铁谱仪智能化检测诊断船舶柴油机故障研究报告[R].南京长江油运公司,南京航空航天大学, 1995.
[42] Jadranka Jakopovic. Approximate Knowledge in LEXIT, An Expert System for Assessing Marine Lubricant Quality and Diagnosing Engine Failures[J]. Computer in Industry, 1991, 17(1):43-47.
[43] Roylance B. J, Albidewi I. A, Laqhari. M. S. Computer-aided Vision Engineering (CAVE)- Quantification of Wear Particle Morphology[J]. Lubrication Engineering, 1994, 50(2):111-116.
[44] Brassard M, Zhang K, Hay D. R. Intelligent debris analyzer: a new tool for monitoring lubrication machines[R]. Storming Meadia Report, No A366743, 1998.
[45]王峰.图像处理技术在铁谱磨粒图像分析中的应用研究[D].武汉理工大学, 2005.
[46]航空发动机设计手册总编委会.传动及润滑系统[M].航空工业出版社, 2002.
[47] Marr D, Hildreth E. Theory of Edge Detection[J]. P roc.R.Soc.Lond.1980, B207:187-217.
[48] Deriche R. Using Canny’s Criteria to Derive a Recursively Implemented Optimal Edge Detection[J].International Journal of Computer Vision, 1987, 1(2):167-187.
[49] Sarkar S, Boyer K. L. Optimal Infinite Impulse Response Edge Detection Filter[J]. IEEE Trans on PAMI, 1991, 13(11):1154-1171.
[50]周凌翔.计算机视觉若干关键问题研究[D].浙江大学, 1997.
[51]杨晖,曲秀杰.图像分割方法综述[J].电脑开发与应用, 2005, 18(3):21-23.
[52] Dwight F, Mix Kraig J. Olejniczak.小波基础及应用教程[M].北京:机械工业出版社,2006.
[53] Sidney Burrus C, Ramesh A, Gopinath.小波与小波变换导论[M].北京:机械工业出版社, 2008.
[54]李弼程,罗建书.小波分析及其应用[M].北京:电子工业出版社, 2003.
[55]孙长峰.基于小波边缘检测的图像去噪方法研究[D].郑州大学, 2007.
[56] Mallat S, Wang W. Singularity Detection and Processing with Wavelet [J]. IEEE Trans. IT, 1992,38(2):617-643.
[57] Mallat S, Zhong S. Characterization of Signals from Multiscale Edges [J]. IEEE Trans. PAMI,1992, 14(7):710-732.
[58]杨冰冰.基于小波分析的图像边缘检测算法研究[D].东南大学, 2005.
[59]周明月,王宏志.基于小波变换的边缘检测[J].长春理工大学学报, 2006, 29(3): 20-22.
[60]郑海疆.小波理论在图像边缘检测种的应用研究[D].厦门大学, 2006.
[61]何晓群,刘文卿.应用回归分析[M].中国人民大学出版社,2001.
[62]刘英华.利用线性回归作预测的研究[D].武汉科技大学, 2007.
[63]纪永凤.灰多元线性回归分析及其应用研究[D].东北师范大学, 2008.
[64]肖建华.智能模式识别方法[M].华南理工大学出版社, 2006.
[65]唐晓初.小波分析及其应用[M].重庆大学出版社, 2006.
[66] Anscombe F. J. Rejection of Outliers[J]. Technometrics, 1960, 2(2):123-147.
[67] Hawkins D. M. Identification of outliers[M]. New York: Chapman and hall, 1980.
[68]胡峰.动态测量数据的合理性检验方法综述.人造卫星观测与研究, 1993, 1:39-54.
[69] Kunsch H. Infinitesmal Robustness for Antoregressive Process[J], Ann Statist. 1984, 12:843-863.
[70] Fox A. J. Outliers in time series[J]. Journal of the Royal Statistical Society. Series B. 1972, 34(3):350-363.
[71] Knorr E. M, Ng R. T. A logarithm for mining distance-based outliers in large datasets[C]. In Proceedings of 24th International Conference of Very Large Data Bases. New York, USA: Morgan Kaufmann Publishers, 1998:392-403.
[72] M. M. Moya, M. W. Koch, L. D. Hostetler. One-class classifier networks for target recognition applications[C]. In Proceedings World Congress on Neural Networks, Portland, 1993:797-801.
[2]陈果.航空器检测与诊断技术导论[M].中国民航出版社, 2007.
[3]杨春生,孟昭荣编译.世界航空安全与事故分析[M].北京:中国民航出版社, 1997.
[4]吴振锋.基于磨粒分析和信息融合的发动机磨损故障诊断技术研究[D].南京航空航天大学, 2001.
[5]范作民,孙春林,白杰.航空发动机故障诊断导论[M].科学出版社, 2004.
[6]陶春虎,习年生,钟培道.航空装备失效典型案例分析[M].北京:国防工业出版社, 1998.
[7] D. P. Anderson著.金元生,杨其明译.磨粒图谱[M].北京:机械工业出版社, 1987.
[8] Tzafestas S. G, Konstantinidis N. I. ENGEXP-an Integrated Environment for The Development and Application of Expert Systems in Equipment and Engine Fault Diagnosis and Repair[J]. Advances in Engineering Software and workstations, 1992, 14(1):3-14.
[9] Raadnui S, Roylance B. J. The classification of wear particle shape[J]. Lubrication Engineering, 1995, 51(5):432-437.
[10] K. K. Yeung, et al. Development of computer-aided image analysis for filter debris analysis[J]. Lubrication Engineering, 1994, 50(4):293-299.
[11] F. T. Barwell. The contribution of particle analysis to the study of wear of metals[J]. Wear, 1983,90:167-181.
[12] A. D. H. Thomas, et al. Computer image analysis for identification of wear Particles[J].Wear, 1991, 142(1):213-226.
[13] Tomasz Pancewicz, Irena Mruk. Holographic contouring for determination of three- dimensional description of surface roughness[J]. Wear, 1996, 199(1):127-131.
[14] T. B. Kirk, et al. Fractal Parameters and computer image analysis applied to wear particles isolated by ferrography[J]. Wear, 1991, 145(2):347-365.
[15] Kun Xu, A. R. Luxmoore. An integrated system for automatic wear particle analysis[J]. Wear, 1997, 208(1):184-193.
[16] Podsiadlo P, Stachowiak G. W. Scale-invariant analysis of wear particle surface morphology[J]. Wear, 2000, 24(2):160-201.
[17] Laghari M. S, Memon Q. A. Knowledge based wear particle analysis[J]. International Journal of Information Technology, 2004, 1(3):91-95.
[18]谢友柏.电磁铁谱仪与铁谱技术的研究[R].西安:西安交通大学科技报告, 1983.
[19]杨志伊.铁谱仪的新发展-KTP旋转式铁谱仪[J].设备管理与维修,1990, 7:22~24.
[20]萧汉梁著.铁谱技术及其在机械监测诊断中的应用[M].北京:人民交通出版社, 1993.
[21]陈国安,葛世荣.基于分形理论的磨合磨损预测模型[J].机械工程学报, 2000, 36(2):29-33.
[22]李忠,曾昭翔.基于BP神经网络的磨损微粒智能识别[J].北方交通大学学报, 1998, 22(1):86-91.
[23]涂群章,左洪福,李艳军.发动机磨损颗粒图像在线检测技术[J].内燃机工程, 2003, 24(5):70-73.
[24]王伟华,殷勇辉.基于径向基函数神经网络的磨粒识别系统[J].摩擦学学报, 2003, 123(4):340-343.
[25]姜旭峰,彭秀华.基于灰色关联分析的磨粒识别[J].润滑与密封, 2005, 171(5):112-116.
[26]高经纬.遗传算法和神经网络在铁谱图像识别中的应用[J].润滑与密封,2004,161(1):16-18.
[27]钟新辉.粗糙集和证据理论在磨粒识别中的应用[J].润滑与密封, 2005, 5:143-145.
[28]康剑莉,芦亚萍,周银生.改进的BP算法在磨粒识别中的应用[J].润滑与密封, 2005, 2:149-154.
[29]李华强,费逸伟.基于Matlab聚类分析的磨粒分类识别研究[J].润滑与密封,2005,3:84-85.
[30]顾大强.基于支持向量机的铁谱磨粒模式识别[J].中国机械工程, 2006, 17(13):1391-1394.
[31]刘国光.基于改进支持向量机的磨粒铁谱识别的研究[J].润滑与密封, 2005, 3:94-98.
[32]左洪福,李立.发动机磨屑尺寸的分布模型及其应用[J].南京航空航天大学学报, 1995, 27(6):784-791.
[33] H. E. Powrie, C. E. Fisher. Engine health monitoring: Towards total prognostics[J]. Aerospace Conference, 1999, 3:11-20.
[34] Powrie H. E. G, McNicholas K. Gas path condition monitoring during Accelerated Mission Testing of a demonstrator engine[C]. The 33rd AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit Seattle. WA, USA: AIAA, 1997: 2904-2914.
[35] Nurse J, Petch C, Fisher C. E. Engine gas path integrity monitoring[C]. Aerotech 94, 1994(7):C470/6/052.
[36]陈果.航空发动机磨损故障的智能融合诊断[J].中国机械工程, 2005, 16(4):299-306.
[37] Tauber T, Johnsen D. Experience with the electronic oil debris monitoring system of the General Electric GE90 gas turbine engine on the Boeing 777 aircraft[J], JOAP International condition monitoring conference, Mobile, Alabama, 2002.
[38] Saba C. S, Wolf J. D, Dell A. J. XRF, a choice method for metal analysis[J], JOAP Internationalcondition monitoring conference, Mobile, Alabama, 2002.
[39] Lee R. J, Toms A. M. An SEM approach to wear debris analysis[J], JOAP International condition monitoring conference, Mobile, Alabama, 2002.
[40] Humphrey G. R, Gorzen D. Prototype instruments built for PRAM[J], JOAP International condition monitoring conference, Mobile, Alabama, 2002.
[41]铁谱仪智能化检测诊断船舶柴油机故障研究报告[R].南京长江油运公司,南京航空航天大学, 1995.
[42] Jadranka Jakopovic. Approximate Knowledge in LEXIT, An Expert System for Assessing Marine Lubricant Quality and Diagnosing Engine Failures[J]. Computer in Industry, 1991, 17(1):43-47.
[43] Roylance B. J, Albidewi I. A, Laqhari. M. S. Computer-aided Vision Engineering (CAVE)- Quantification of Wear Particle Morphology[J]. Lubrication Engineering, 1994, 50(2):111-116.
[44] Brassard M, Zhang K, Hay D. R. Intelligent debris analyzer: a new tool for monitoring lubrication machines[R]. Storming Meadia Report, No A366743, 1998.
[45]王峰.图像处理技术在铁谱磨粒图像分析中的应用研究[D].武汉理工大学, 2005.
[46]航空发动机设计手册总编委会.传动及润滑系统[M].航空工业出版社, 2002.
[47] Marr D, Hildreth E. Theory of Edge Detection[J]. P roc.R.Soc.Lond.1980, B207:187-217.
[48] Deriche R. Using Canny’s Criteria to Derive a Recursively Implemented Optimal Edge Detection[J].International Journal of Computer Vision, 1987, 1(2):167-187.
[49] Sarkar S, Boyer K. L. Optimal Infinite Impulse Response Edge Detection Filter[J]. IEEE Trans on PAMI, 1991, 13(11):1154-1171.
[50]周凌翔.计算机视觉若干关键问题研究[D].浙江大学, 1997.
[51]杨晖,曲秀杰.图像分割方法综述[J].电脑开发与应用, 2005, 18(3):21-23.
[52] Dwight F, Mix Kraig J. Olejniczak.小波基础及应用教程[M].北京:机械工业出版社,2006.
[53] Sidney Burrus C, Ramesh A, Gopinath.小波与小波变换导论[M].北京:机械工业出版社, 2008.
[54]李弼程,罗建书.小波分析及其应用[M].北京:电子工业出版社, 2003.
[55]孙长峰.基于小波边缘检测的图像去噪方法研究[D].郑州大学, 2007.
[56] Mallat S, Wang W. Singularity Detection and Processing with Wavelet [J]. IEEE Trans. IT, 1992,38(2):617-643.
[57] Mallat S, Zhong S. Characterization of Signals from Multiscale Edges [J]. IEEE Trans. PAMI,1992, 14(7):710-732.
[58]杨冰冰.基于小波分析的图像边缘检测算法研究[D].东南大学, 2005.
[59]周明月,王宏志.基于小波变换的边缘检测[J].长春理工大学学报, 2006, 29(3): 20-22.
[60]郑海疆.小波理论在图像边缘检测种的应用研究[D].厦门大学, 2006.
[61]何晓群,刘文卿.应用回归分析[M].中国人民大学出版社,2001.
[62]刘英华.利用线性回归作预测的研究[D].武汉科技大学, 2007.
[63]纪永凤.灰多元线性回归分析及其应用研究[D].东北师范大学, 2008.
[64]肖建华.智能模式识别方法[M].华南理工大学出版社, 2006.
[65]唐晓初.小波分析及其应用[M].重庆大学出版社, 2006.
[66] Anscombe F. J. Rejection of Outliers[J]. Technometrics, 1960, 2(2):123-147.
[67] Hawkins D. M. Identification of outliers[M]. New York: Chapman and hall, 1980.
[68]胡峰.动态测量数据的合理性检验方法综述.人造卫星观测与研究, 1993, 1:39-54.
[69] Kunsch H. Infinitesmal Robustness for Antoregressive Process[J], Ann Statist. 1984, 12:843-863.
[70] Fox A. J. Outliers in time series[J]. Journal of the Royal Statistical Society. Series B. 1972, 34(3):350-363.
[71] Knorr E. M, Ng R. T. A logarithm for mining distance-based outliers in large datasets[C]. In Proceedings of 24th International Conference of Very Large Data Bases. New York, USA: Morgan Kaufmann Publishers, 1998:392-403.
[72] M. M. Moya, M. W. Koch, L. D. Hostetler. One-class classifier networks for target recognition applications[C]. In Proceedings World Congress on Neural Networks, Portland, 1993:797-801.