基于刀具检测仪的数字图像自动聚焦技术研究
|
摘要
随着图像处理技术与计算机技术及检测技术的结合,形成的视觉检测系统可以完成对微小尺寸的高精度测量,是其它设备所无法完成的。而自动聚焦是视觉检测系统中的一项关键,是图像处理、模式识别、精密定位等后续操作的基础。只有快速、精确地采集到清晰的图片,才可以有效地完成后续检测工作。
本课题以刀具检测仪为研究平台,围绕着自动聚焦技术研究了图像的清晰度评价函数、焦点搜索算法和聚焦区域选择等相关技术,得出适用于刀具图像的函数算法。主要包括以下几个方面：
1)了解课题的来源和背景,掌握自动聚焦技术的国内外发展状况和关键技术,并对与自动聚焦技术相关的理论原理作了详细的解释说明。
2)分析了自动聚焦算法的相关理论,包括三个模块：清晰度评价函数、聚焦搜索策略算法和聚焦区域选择的影响。清晰度评价函数是对刀具图像的离焦程度进行评价,它是有参考的图像质量评价,本文对常用清晰度评价函数做了详细的分析比较,研究结果表明8邻域拉普拉斯算子在处理刀具图像时精度高；聚焦搜索策略算法的目的是实现聚焦点的搜索和定位,快速找到精确聚焦位置。本文分析了常用的焦点搜索算法,对算法的流程进行了说明,结果表明对爬山算法改进后的“三点判别法”避免了局部峰值造成的误判,精度高；聚焦区域选择是选择感兴趣区域进行聚焦,可以减少计算量。本文分析了各区域选择算法的优缺点,将其与整幅图像的调焦曲线进行对比分析,结果表明多点取窗区域选择算法效果好,计算效率高,避免了对整幅图像进行计算时的数据量大、时间长等问题。
With the development of image processing compared with computer technology and detection technology, formed the visual detection system. Use the visual detection system we can measure the small size goods with a high precision. But if we use other equipments we can not finish the precision measurement. Auto focus is the key technology in visual detection system, and auto focus is the foundation of image processing, pattern recognition, precise positioning and so on follow-up operations. Only quick and accurate acquire the clear image, we can complete the subsequent work effectively.
In this thesis we take the visual tool detector as the research platform, around the technology of auto focus, research on the technology of clarity-evaluation function, focus search algorithm and region selection. Then we get the better algorithms that are suitable for the tool images, which include the following parts:
1) In this paper, the source of the subject and background are described in detail and the development of auto-focus on the current status is introduced. The paper also gives a detailed introduction of basic theory principles which are related to the subject.
2) Analysis the relevant theory of auto focus algorithm, including three modules:the clarity-evaluation function, focus search strategy algorithm and the effection of the region selection to the clarity-evaluation function. The clarity-evaluation function is evaluated for the tool image, judge the image is focus or defocus. It is a reference to the image quality. In this paper, we analysis and compare the common clarity-evaluation functions. The results show that the Laplace (8 neighborhood) has a higher precision in the tool image processing. The purpose of focus search algorithm is to achieve the searching and positioning of focus. Then quickly find the accurate focus position. In this thesis we analyze the common focus search algorithms, describe the flow chart of each algorithms. The results show that the "three point method" which improves from the hill climbing algorithm can avoid the local peak caused by error and this method has a high precision. The purpose of the region selection is to select the interested regions. Only analysis the interested regions, we can reduce the amount of computation. This thesis analyzes the advantages and disadvantages of each region selection algorithm, and then compared with the whole image. The results show that the region selection of more point window has a good effect, high computational efficiency. If we analysis the whole image, the data is large and the time is long, but if we use this method, it can reduce the amount of data, improve the efficiency.
本课题以刀具检测仪为研究平台,围绕着自动聚焦技术研究了图像的清晰度评价函数、焦点搜索算法和聚焦区域选择等相关技术,得出适用于刀具图像的函数算法。主要包括以下几个方面：
1)了解课题的来源和背景,掌握自动聚焦技术的国内外发展状况和关键技术,并对与自动聚焦技术相关的理论原理作了详细的解释说明。
2)分析了自动聚焦算法的相关理论,包括三个模块：清晰度评价函数、聚焦搜索策略算法和聚焦区域选择的影响。清晰度评价函数是对刀具图像的离焦程度进行评价,它是有参考的图像质量评价,本文对常用清晰度评价函数做了详细的分析比较,研究结果表明8邻域拉普拉斯算子在处理刀具图像时精度高；聚焦搜索策略算法的目的是实现聚焦点的搜索和定位,快速找到精确聚焦位置。本文分析了常用的焦点搜索算法,对算法的流程进行了说明,结果表明对爬山算法改进后的“三点判别法”避免了局部峰值造成的误判,精度高；聚焦区域选择是选择感兴趣区域进行聚焦,可以减少计算量。本文分析了各区域选择算法的优缺点,将其与整幅图像的调焦曲线进行对比分析,结果表明多点取窗区域选择算法效果好,计算效率高,避免了对整幅图像进行计算时的数据量大、时间长等问题。
With the development of image processing compared with computer technology and detection technology, formed the visual detection system. Use the visual detection system we can measure the small size goods with a high precision. But if we use other equipments we can not finish the precision measurement. Auto focus is the key technology in visual detection system, and auto focus is the foundation of image processing, pattern recognition, precise positioning and so on follow-up operations. Only quick and accurate acquire the clear image, we can complete the subsequent work effectively.
In this thesis we take the visual tool detector as the research platform, around the technology of auto focus, research on the technology of clarity-evaluation function, focus search algorithm and region selection. Then we get the better algorithms that are suitable for the tool images, which include the following parts:
1) In this paper, the source of the subject and background are described in detail and the development of auto-focus on the current status is introduced. The paper also gives a detailed introduction of basic theory principles which are related to the subject.
2) Analysis the relevant theory of auto focus algorithm, including three modules:the clarity-evaluation function, focus search strategy algorithm and the effection of the region selection to the clarity-evaluation function. The clarity-evaluation function is evaluated for the tool image, judge the image is focus or defocus. It is a reference to the image quality. In this paper, we analysis and compare the common clarity-evaluation functions. The results show that the Laplace (8 neighborhood) has a higher precision in the tool image processing. The purpose of focus search algorithm is to achieve the searching and positioning of focus. Then quickly find the accurate focus position. In this thesis we analyze the common focus search algorithms, describe the flow chart of each algorithms. The results show that the "three point method" which improves from the hill climbing algorithm can avoid the local peak caused by error and this method has a high precision. The purpose of the region selection is to select the interested regions. Only analysis the interested regions, we can reduce the amount of computation. This thesis analyzes the advantages and disadvantages of each region selection algorithm, and then compared with the whole image. The results show that the region selection of more point window has a good effect, high computational efficiency. If we analysis the whole image, the data is large and the time is long, but if we use this method, it can reduce the amount of data, improve the efficiency.
引文
[1]吴世法.近代成像技术与图像处理[M].北京：国防工业出版社,1997.3.
[2]沈庭芳,方子文编著.数字图像处理及模式识别[M].北京：北京理工大学出版社,1998.7.
[3]夏德深,傅德盛.现代图像处理技术与应用[M].南京：东南大学出版社,1997.5.
[4]Kenneth. R. Castleman著,朱志刚等译.数字图像处理[M].北京：电子工业出版社,2002.2.
[5]章毓晋.图像工程(中册)一图像分析[M].北京：清华大学出版社,2005.1.
[6]王向军,王峰.显微精密成像与微型机械尺寸检测技术[J].光学精密工程,2001,9(6)：511-513.
[7]Guo Ling-hua, Ding Xiao-qing, Fang Mao-fa. Using the minimum entropy as a criterion for the in-orbit automatic focusing of space solar telescope, Optical Technique, 2003,4:391-393.
[8]周敬.刀具预调仪图像测量软件系统开发[D].太原：中北大学机械制造及其自动化专业,2011.
[9]刘力双.电子摄像式刀具预调测量仪的研究[D].天津：天津大学精密仪器与光电子工程学院,2006.
[10]王虎.自动对焦原理及方法[J].科技信息,2008,13：38-39.
[11]林小倩,林斌,潘泰才.基于CMOS单点激光三角法测距系统设计[J].光学仪器,2006,28(2)：27-30.
[12]卜坤亮,王建东,樊玮虹等.基于FPGA芯片的超声波红外线测距系统[J].微处理机,2006,19(2)：88-90.
[13]瞿蓬,林喜荣.一种基于图像处理的自动调焦系统[J].电子技术应用,2002,21(10)：33-35.
[14]于谦.光学检查镜头的MTF测试方法研究[D].杭州：浙江大学光电信息工程学系,2010.
[15]原育凯.光学系统的自动调焦方法[J].红外.2004,15-21.
[16]Gopal Surya, Murali Subbarao. Depth from defocus by Changing camera aperture:a spatial domain approach[J]. Proceeding of IEEE Computer Society Conference on Computer vision and Pattern Recognition,1993,61-67.
[17]裴锡宇,冯华君,李奇等.一种基于频谱分析的离焦深度自动对焦法[J].光电工程,2003,30(5)：62-65.
[18]孙明磊,宗光华,余志伟等.基于图像分析的显微视觉自动聚焦系统[J].北京航空航天大学学报.2005,31(2)：193-198.
[19]刘煜.高分辨率CCD 卫星自动调焦的研究[D].哈尔滨：哈尔滨工业大学电子与光电工程系,1993.
[20]TTE Yeo, S H Ong, Jayasooriah and R Sinniah. Autofocusing for Tissue Microscope[J]. Image and Vision Computing,1993,11 (10):629-639.
[21]Jarvis R.A. IFocus Optimisation Criteria for Computer Image Processing[J]. Microscope, 1976,(24):163-180.
[22]薛实福,李庆祥.精密仪器设计[J].清华大学出版社,1991：225-244.
[23]Ren C. Luo. Issues and Approaches of Automatic Focusing Algorithms for Intelligent Robot Eye-in-hand System[J].Journal of Robotic System,1987(4):459-476
[24]Firestone, Cook K, Culp K, Talsania N, Preston K Jr. Comparison of Autofocus Methods for Automated Microscopy[J]. Cytometry,1991,12(3):195-206.
[25]白立芬,李庆祥.基于图像处理的显微镜自动调焦方法研究[J].仪器仪表学报,1999,20(6)：612-614.
[26]陈加卿,高永毅,黄锡鹏等.工业电视摄像机的微机控制[J].电视技术,1985,(5)：11-15.
[27]钱恒迈.光刻机自动调焦装置[J].光学工程,1986,(5)：24-29.
[28]白立芬,于水等.基于图像处理的显微镜自动对焦技术研究[J].仪器仪表学报,1999,20(6)：613-617.
[29]祝世平,房建成,周锐.一种新的能量谱——熵图像聚焦评价函数[J].北京航空航天大学学报,1999,25(6)：720-723.
[30]任四刚.图像式自动聚焦系统及其测量技术的研究[D].重庆：重庆大学光电工程学院,2002.
[31]郑玉珍,吴勇,倪旭祥.实时自动对焦的研究[J].光电工程,2004,31(4)：64-66.
[32]周贤,姜威,朱孔凤.基于小波多分辨率分析的自动聚焦算法[J].计算机应用与软件,2006,23(12)：104-105,134.
[33]张辉.CCD图像检测系统自动调焦的研究[D].哈尔滨：哈尔滨工业大学机电工程学院,2007.
[34]张彭舜.基于数字图像的自动聚焦技术研究[D].哈尔滨：哈尔滨工业大学光学工程学院,2008.
[35]樊攀.基于数字图像处理的自动聚焦系统算法研究[D].哈尔滨：哈尔滨工业大学光学工程学院,2009.
[36]余超,王伯雄,郑汉卿等.显微镜自动粗调焦的TenenGrad改进算法[J].光学精密工程,2007,15(5)：784-790.
[37]刘焕雨,熊文卓,万秋华等.基于图像处理方法的自动调焦系统的研制[J].测试技术学报,2007,21(1)：13-17.
[38]蒋婷.基于图像处理的自动对焦理论和技术研究[D].武汉：武汉理工大学机电工程学院,2008.
[39]Wang GuangLin, Chen DaWei. Research on CCD-based microscopic measuring system for
[40]黄其昆.自动调焦二十年[J].大众摄影,2001,(5)：52-55.geometry error of high precision part. SPIE,1998,3558:175-185.
著,阮秋琦,阮宇智等译.数字图像处理(第二
[41]Rafael C.Gonzalez, Richard E.Woods版)[M].北京：电子工业出版社,2007.8.[42]张佳平,黄元庆.数字图像刀具几何参数在线测量系统的研究[J].中国仪
表,2007,(2)：48-50.[43]侯学智.数字图像刀具测量机[D].成都：电子科技大学机械电子工程学院,2004.
[50]康宗明,张利,谢攀等.一种基于能量和熵的自动聚焦算法[J].电子学报,2003,
[44]Tabatabai A J, Mitchell O R. Edge Location to Sub-pixel Values in Digital Imagery[J].IEEE Trans on PAMI,1984,6(2):188-201.
[45]Lyvers E P, Mitchell O R. Sub-pixel Measurements Using a Moment-Based Edge Operator[J].IEEE Trans on PAMI,1989,11(2):1293-1309.
[46]FransC, Agroen, IanT. Young, Guido Ligthart. A Comparison of Different Focus Functions for Use in Autofocus Algorithms[J].Cytometry,6:81-91.
[47]H. Harms, H. M. Aus. Compaison of Digital Focus Criteria for a TV Microscope Systeml[J].Cytometry,1984:236-243.
[48]Brenner J F, Dew B S, Horton J B, King T, Neirath P W, Sellers W D. An Automated Microscope for Cytologic Research[J].1971,(24):100-110.
[49]Mason D C, Green D K. Automated Focusing of a Computer Controlled Microscopel[J]. IEEE Trans Biomed Eng BME,1975,(122):137-150.
31(4)：552-555.[51]黄剑琪,冯华君,徐之海等.边缘特征增强算法和小波分析在精确调焦中的应用[J].光
子学报,2000,29(10)：932-934.[56]周贤,姜威等.基于边缘能量的自动聚焦算法[J].光学技术,2006,32(1)：36-40.
[52]J. Cvaiedes, S. Gubruz. No-reference Sharpness Mertric Based on Local Edge Kurtosis[J].Intenrational Conference on Image Processing,2002,3:53-56.
[53]L. Dalong, R. M. Mersereau, S. Simske. Blur Identification Based on Kurtosis Minimization[J].IEEE Intenrational Conference on Image Processing,2005,1:905-908.
[54]Rafael C, Gonzalez, Richard E, Woods. Digital Image Processing(Second Edition) [M]. Publishing House of Electronics Industry,2008:105-108.
[55]Marr D. Vision, W.H. Freeman Company[J].1982:78-91.
[57]H. Harms, H. M. Aus. Comparison of Digital Focus Criteria for a TV Microscope Systeml[J].Cytometry,1984:236-243.
[58]Brenner, J F, Dew, B S, Horton, J B, King, T, Neirath, P W and Sellers, W D. An Automated Microscope for Cytologic Research[J].1971,(24):100-110.
[59]Mason D C, Green D K. Automated Focusing of a Computer Controlled Microscopel[J].IEEE Trans Biomed Eng BME,1975,(122):137-150.
[60]Mendelsohn ML, Mayall BH. Computer-oriented Analysis of Human Chromosomesl[J]. Comput Biol Med,1992,(2):137-150.
[611 Charistoper F. Batten. Auto-focusing and Astigmatism Correction in theScanning Electron Microscope[J]. University of Cambridge,2000,20-60.
[62]Kang-Sun Choi, Jun-Suk Lee, Sung-Jaeko. New Auto-focusing TechniqueUsing the Frequency Selective Weighted Median Filter for Video Cameras[J]. IEEE Transactions on Consumer Electron,1999,45(3):520-827.
[63]袁珂.数字成像系统的自动调焦理论和技术研究[D].长沙：长沙理工大学计算机与通信工程学院,2006.
[2]沈庭芳,方子文编著.数字图像处理及模式识别[M].北京：北京理工大学出版社,1998.7.
[3]夏德深,傅德盛.现代图像处理技术与应用[M].南京：东南大学出版社,1997.5.
[4]Kenneth. R. Castleman著,朱志刚等译.数字图像处理[M].北京：电子工业出版社,2002.2.
[5]章毓晋.图像工程(中册)一图像分析[M].北京：清华大学出版社,2005.1.
[6]王向军,王峰.显微精密成像与微型机械尺寸检测技术[J].光学精密工程,2001,9(6)：511-513.
[7]Guo Ling-hua, Ding Xiao-qing, Fang Mao-fa. Using the minimum entropy as a criterion for the in-orbit automatic focusing of space solar telescope, Optical Technique, 2003,4:391-393.
[8]周敬.刀具预调仪图像测量软件系统开发[D].太原：中北大学机械制造及其自动化专业,2011.
[9]刘力双.电子摄像式刀具预调测量仪的研究[D].天津：天津大学精密仪器与光电子工程学院,2006.
[10]王虎.自动对焦原理及方法[J].科技信息,2008,13：38-39.
[11]林小倩,林斌,潘泰才.基于CMOS单点激光三角法测距系统设计[J].光学仪器,2006,28(2)：27-30.
[12]卜坤亮,王建东,樊玮虹等.基于FPGA芯片的超声波红外线测距系统[J].微处理机,2006,19(2)：88-90.
[13]瞿蓬,林喜荣.一种基于图像处理的自动调焦系统[J].电子技术应用,2002,21(10)：33-35.
[14]于谦.光学检查镜头的MTF测试方法研究[D].杭州：浙江大学光电信息工程学系,2010.
[15]原育凯.光学系统的自动调焦方法[J].红外.2004,15-21.
[16]Gopal Surya, Murali Subbarao. Depth from defocus by Changing camera aperture:a spatial domain approach[J]. Proceeding of IEEE Computer Society Conference on Computer vision and Pattern Recognition,1993,61-67.
[17]裴锡宇,冯华君,李奇等.一种基于频谱分析的离焦深度自动对焦法[J].光电工程,2003,30(5)：62-65.
[18]孙明磊,宗光华,余志伟等.基于图像分析的显微视觉自动聚焦系统[J].北京航空航天大学学报.2005,31(2)：193-198.
[19]刘煜.高分辨率CCD 卫星自动调焦的研究[D].哈尔滨：哈尔滨工业大学电子与光电工程系,1993.
[20]TTE Yeo, S H Ong, Jayasooriah and R Sinniah. Autofocusing for Tissue Microscope[J]. Image and Vision Computing,1993,11 (10):629-639.
[21]Jarvis R.A. IFocus Optimisation Criteria for Computer Image Processing[J]. Microscope, 1976,(24):163-180.
[22]薛实福,李庆祥.精密仪器设计[J].清华大学出版社,1991：225-244.
[23]Ren C. Luo. Issues and Approaches of Automatic Focusing Algorithms for Intelligent Robot Eye-in-hand System[J].Journal of Robotic System,1987(4):459-476
[24]Firestone, Cook K, Culp K, Talsania N, Preston K Jr. Comparison of Autofocus Methods for Automated Microscopy[J]. Cytometry,1991,12(3):195-206.
[25]白立芬,李庆祥.基于图像处理的显微镜自动调焦方法研究[J].仪器仪表学报,1999,20(6)：612-614.
[26]陈加卿,高永毅,黄锡鹏等.工业电视摄像机的微机控制[J].电视技术,1985,(5)：11-15.
[27]钱恒迈.光刻机自动调焦装置[J].光学工程,1986,(5)：24-29.
[28]白立芬,于水等.基于图像处理的显微镜自动对焦技术研究[J].仪器仪表学报,1999,20(6)：613-617.
[29]祝世平,房建成,周锐.一种新的能量谱——熵图像聚焦评价函数[J].北京航空航天大学学报,1999,25(6)：720-723.
[30]任四刚.图像式自动聚焦系统及其测量技术的研究[D].重庆：重庆大学光电工程学院,2002.
[31]郑玉珍,吴勇,倪旭祥.实时自动对焦的研究[J].光电工程,2004,31(4)：64-66.
[32]周贤,姜威,朱孔凤.基于小波多分辨率分析的自动聚焦算法[J].计算机应用与软件,2006,23(12)：104-105,134.
[33]张辉.CCD图像检测系统自动调焦的研究[D].哈尔滨：哈尔滨工业大学机电工程学院,2007.
[34]张彭舜.基于数字图像的自动聚焦技术研究[D].哈尔滨：哈尔滨工业大学光学工程学院,2008.
[35]樊攀.基于数字图像处理的自动聚焦系统算法研究[D].哈尔滨：哈尔滨工业大学光学工程学院,2009.
[36]余超,王伯雄,郑汉卿等.显微镜自动粗调焦的TenenGrad改进算法[J].光学精密工程,2007,15(5)：784-790.
[37]刘焕雨,熊文卓,万秋华等.基于图像处理方法的自动调焦系统的研制[J].测试技术学报,2007,21(1)：13-17.
[38]蒋婷.基于图像处理的自动对焦理论和技术研究[D].武汉：武汉理工大学机电工程学院,2008.
[39]Wang GuangLin, Chen DaWei. Research on CCD-based microscopic measuring system for
[40]黄其昆.自动调焦二十年[J].大众摄影,2001,(5)：52-55.geometry error of high precision part. SPIE,1998,3558:175-185.
著,阮秋琦,阮宇智等译.数字图像处理(第二
[41]Rafael C.Gonzalez, Richard E.Woods版)[M].北京：电子工业出版社,2007.8.[42]张佳平,黄元庆.数字图像刀具几何参数在线测量系统的研究[J].中国仪
表,2007,(2)：48-50.[43]侯学智.数字图像刀具测量机[D].成都：电子科技大学机械电子工程学院,2004.
[50]康宗明,张利,谢攀等.一种基于能量和熵的自动聚焦算法[J].电子学报,2003,
[44]Tabatabai A J, Mitchell O R. Edge Location to Sub-pixel Values in Digital Imagery[J].IEEE Trans on PAMI,1984,6(2):188-201.
[45]Lyvers E P, Mitchell O R. Sub-pixel Measurements Using a Moment-Based Edge Operator[J].IEEE Trans on PAMI,1989,11(2):1293-1309.
[46]FransC, Agroen, IanT. Young, Guido Ligthart. A Comparison of Different Focus Functions for Use in Autofocus Algorithms[J].Cytometry,6:81-91.
[47]H. Harms, H. M. Aus. Compaison of Digital Focus Criteria for a TV Microscope Systeml[J].Cytometry,1984:236-243.
[48]Brenner J F, Dew B S, Horton J B, King T, Neirath P W, Sellers W D. An Automated Microscope for Cytologic Research[J].1971,(24):100-110.
[49]Mason D C, Green D K. Automated Focusing of a Computer Controlled Microscopel[J]. IEEE Trans Biomed Eng BME,1975,(122):137-150.
31(4)：552-555.[51]黄剑琪,冯华君,徐之海等.边缘特征增强算法和小波分析在精确调焦中的应用[J].光
子学报,2000,29(10)：932-934.[56]周贤,姜威等.基于边缘能量的自动聚焦算法[J].光学技术,2006,32(1)：36-40.
[52]J. Cvaiedes, S. Gubruz. No-reference Sharpness Mertric Based on Local Edge Kurtosis[J].Intenrational Conference on Image Processing,2002,3:53-56.
[53]L. Dalong, R. M. Mersereau, S. Simske. Blur Identification Based on Kurtosis Minimization[J].IEEE Intenrational Conference on Image Processing,2005,1:905-908.
[54]Rafael C, Gonzalez, Richard E, Woods. Digital Image Processing(Second Edition) [M]. Publishing House of Electronics Industry,2008:105-108.
[55]Marr D. Vision, W.H. Freeman Company[J].1982:78-91.
[57]H. Harms, H. M. Aus. Comparison of Digital Focus Criteria for a TV Microscope Systeml[J].Cytometry,1984:236-243.
[58]Brenner, J F, Dew, B S, Horton, J B, King, T, Neirath, P W and Sellers, W D. An Automated Microscope for Cytologic Research[J].1971,(24):100-110.
[59]Mason D C, Green D K. Automated Focusing of a Computer Controlled Microscopel[J].IEEE Trans Biomed Eng BME,1975,(122):137-150.
[60]Mendelsohn ML, Mayall BH. Computer-oriented Analysis of Human Chromosomesl[J]. Comput Biol Med,1992,(2):137-150.
[611 Charistoper F. Batten. Auto-focusing and Astigmatism Correction in theScanning Electron Microscope[J]. University of Cambridge,2000,20-60.
[62]Kang-Sun Choi, Jun-Suk Lee, Sung-Jaeko. New Auto-focusing TechniqueUsing the Frequency Selective Weighted Median Filter for Video Cameras[J]. IEEE Transactions on Consumer Electron,1999,45(3):520-827.
[63]袁珂.数字成像系统的自动调焦理论和技术研究[D].长沙：长沙理工大学计算机与通信工程学院,2006.