保健酒智能视觉检测机器人技术研究
|
摘要
我国有着巨大的饮料酒生产市场以及饮料酒产品的高增长率，但饮料酒生产线的自动化装备水平亟待提高，尤其在饮料酒生产线上的检测环节，主要依赖人工检测。近年来，随着消费水平的提高和相关法律、法规的完善，在饮料酒类行业中，产品质量变得越来越重要。人工检测效率低、精度低、漏检误检率高，面对国外先进成套饮料酒生产装备对国内市场的冲击，研究具有自主知识产权的饮料酒自动检测设备具有十分重要的价值。
本论文旨在研制面向大型高速保健酒自动化生产线上的产品检测智能机器人。保健酒智能视觉检测机器人是以多传感器信息融合为基础，利用机器视觉和智能控制技术，在不影响正常的安全生产过程的同时，可实现高速保健酒自动化生产线上的自动检测，并能把不合格保健酒实时自动分离出生产线。
保健酒智能视觉检测机器人涉及机器视觉、机械设计与传动、光机电一体化、传感器检测与信息融合、光学成像、数字图像处理和通信等多学科领域技术。本文从饮料酒检测现状出发，首先分析了机器人实施的技术路线和整体设计方案，研究了机械结构、机电控制系统、光学成像与视觉信息获取，详细设计开发了针对保健酒液中的可见异物检测和外观检测算法，并经过了大量实验论证，最后研制出保健酒智能视觉检测机器人样机，并把相关技术推广到大输液和口服液等医药自动化生产线上，为饮料医药行业的产品安全生产提供重要保障。归纳起来，本文主要工作和取得的创新成果有：
1、阐述了饮料酒产品的检测现状，指出研制智能视觉检测机器人的重要意义，概述了机器视觉技术，分析了保健酒中可见异物及外观缺陷的成因，重点分析了国内外饮料酒检测的现状。
2、对保健酒智能视觉检测机器人系统进行了可行性分析论证，给出了整体实施技术路线和技术方案，设计了直线间歇式、圆盘固定位置拍照式和旋转跟踪式三种保健酒智能视觉检测机器人的机械结构，并进行了对比分析，研究了基于多IPC和PLC的分布式精密光机电控制结构，分析了伺服跟踪拍摄控制的“加速跟上-同角速度运转-快速返回”三个阶段过程，提出了黑色异物和装量检测工位、白色异物检测工位和瓶盖检测工位的光学成像方法，给出了机器人的检测流程，对比研究了多种次品分拣方案，分析了机器人的软件构架。
3、对保健酒中的可见异物对象进行了运动学分析，并把可见异物和气泡、瓶壁刮痕刻度等进行了对比，对基本滤波算法进行了分析研究了一种改进型加权滤波处理算法，并提出了基于中值的加权均值滤波算法，既能有效的滤除噪声，又能有效的保留运动目标，根据保健酒特定的瓶型进行了检测区域的标定，提出了一种基于边缘统计的经验规则定位法进行图像定位，并得到了良好的定位效果。
4、研究了保健酒液序列图像中可见异物的分割与识别方法。分析了背景减除的运动目标提取方法，详细介绍了基于时间差分的运行目标提取，研究了基于Fisher评价函数的阈值分割，详细分析了基于改进型Mean Shift和Kalman滤波的运动目标跟踪方法，并进行了实例分析与测试，从运动目标的位置、形状和运动轨迹等方面，构建支持向量，研究了基于支持向量机的运动目标识别，进行了图像标定，然后从压像素角度分析，对可见异物进行测量，并判断异物尺寸是否在允许范围内，通过实验，验证了可见异物分割与识别算法的可行性和有效性。
5、研究了保健酒产品的外观检测算法。针对侧瓶盖缺陷，提出了多重扫描的边缘特殊点重心法定位，针对顶面瓶盖，提出了一种多重搜索定位法定位，得到定位点后，研究了一种基于分层思想的瓶盖缺陷图像快速匹配检测方法实现瓶盖检测，针对Sobel算子边缘检测的优点和缺陷，提出了基于改进型Sobel算子，得到液面线，并根据经验规则判断液面高度，实现装量的检测。
6、实现了保健酒智能视觉检测机器人的样机研制及推广应用。给出了机器人的实物图和各关键部分装置图，从人机界面、测试界面、控制软件和数据库模块四个方面分析了软件部分的构成，最后分析了保健酒智能视觉检测机器人的Knapp-Kushner测试方法，给出了测试数据，最后把保健酒智能视觉检测机器人的相关技术推广到大输液视觉检测机器人、安瓿视觉检测机器人和口服液视觉检测机器人等设备上，并得到了实际应用。
在保健酒智能视觉检测机器人研发过程中，从方案设计到测试实验，从图像预处理与定位算法，到可见异物检测与识别和外观检测方法的研究，从机械设计到精密光机电控制与传动，从样机研制到推广应用，发现并解决了大量的理论和实际问题，为设备的应用及后一步的升级开发提供了宝贵的经验，给相关设备的研发带来了良好的参考价值，为实现饮料酒和医药自动化生产线上的产品质量安全提供了保障。
China has a large alcoholic drinks market and their production increased quicklyyear by year, however the automation level of the equipment, especially theautomatic detection, in the alcoholic drinks production line needs improving. Inrecent years, with the increase of consumption and perfection of laws and regulations,the product quality is more and more important in the alcoholic drinks industry. Thenaked-eye detection, which is not efficient and accurate enough, is being challengedby the advanced alcoholic drinks production equipments from foreign countries.Therefore, it is significant to research on alcoholic detection equipment withproprietary intellectual property right in China.
The project is designed to develop an intelligent robot to detect the products inthe large-scale and super-speed health liquor automatic production line. Based onmulti-sensor information fusion detection and recognition, the intelligent robot candetect the products automatically and separate the unqualified products from theproduction line. Meanwhile, the intelligent robot can be applied in other fields, suchas drinks and medicine industries, and insure the safety of products in such fields.
Health liquor intelligent visual detection robot involves technologies in manydisciplines, such as machine vision, machine design and transmission, electricalcontrol, sensor detection and information integration, optical imaging, digital imageprocessing, communication and so on. This paper, firstly, analyzes the overall designproposal and the technologies that the intelligent robot involves. It secondly presentsthe proved algorithm aimed at detection of visible foreign matter in the health liquorand the bottle detection by researching on the mechanical structure, electrical controlsystem, optical imaging, and visual information acquisition. Finally, this paperpresents a prototype of the health liquor intelligent visual detection robot and appliesthe technology into the production of some medicine, such as transfusion and oralliquid. All in all, the innovative achievements presented in this paper are as follows.
1. This paper elaborates the present situation of alcoholic drinks detection,points out the significance of developing an intelligent visual detection robot,introduces machine vision technology, analyzes the reasons why the visible foreignmatter in the health liquor and the bottle defect occur, and focuses on the presentsituation of liquor in China and foreign countries.
2. This paper proves the feasibility of the health liquor intelligent visualdetection robot system, presents the overall technical proposal, compares three kindsof mechanical structure of health liquor intelligent visual detection robot: linearintermittence, taking pictures in a fixed position and rotated tracking, researchesdistributed electrical control structure based on IPC and PLC, analyzes three stagesphotographed by tracking: acceleration, revolve at the same angular speed, rapidreturn, designs the optical imaging that involves detection posts of black foreignmatter, capacity, white foreign matter and bottle cap, presents the robot’s detectionprocedures, compares many kinds of unqualified products sorting methods andanalyzes robot’s software framework.
3. This paper analyzes kinematically the visible foreign matter in the healthliquor, compares visible foreign matter, air bubbles, scratches on the bottle walls, putforwards an improved weighted filter algorithm on the basis of analysis of basicfiltering algorithm, promotes an improved adaptive-filtering algorithm based on theextremum and mean value, which can not only filter efficiently the noise, but alsoremain the moving targets, by researching on the weighted average filteringalgorithm based on mid-value, proposes empirical rule positioning method to positionimages based on edge statistics, which achieves good positioning effect.
4. This paper researches on method of segmentation and identification of visibleforeign matter in the sequential pictures of health liquor, analyzes method ofextraction of moving target with background subtraction, researches on thresholdsegmentation based on Fisher evaluation function, analyzes moving target trackingmethod based on improved Mean Shift and Kalman filtering. This paper also presentsthe living examples to analyze and test. Based on the shape of the moving target,such as length, width, rectangular degree and degree of circularity, as well asestimation of movement, this paper presents support vectors, researches on movingtarget identification on the basis of support vectors, conducts image calibration. Thenanalyzing from the pressure pixel, this paper presents the measurement of visibleforeign matter and the judgment whether the foreign matter size is within the allowedlimits. Finally, this paper presents the feasibility of visible foreign mattersegmentation and identification algorithm through experiments.
5. This paper researches on liquor bottle detection algorithm. Aimed at thedefect that the bottle cape inclines at one side, this paper proposes positioning in theedge special point based on center of gravity and multiple scanning. Directing atdefect that the bottle cap may emboss, this paper promotes a positioning method based on multiple searching. It also researches on detection method of bottle capdefect pictures quickly matching based on layering once the positioning point iscertain. Aimed at the advantages and disadvantages of Sobel operator edge detection,this paper presents an improved Sobel operator to acquire the liquid capacity line andrealize the capacity detection on the basis of liquid surface height judged byexperience.
6. This paper presents the prototype of health liquor intelligent visual detectionrobot and promotes its application. It shows the picture of the robot and set-updiagram of the key components, analyzes the software constitution from four aspectsthat is from human-computer interface, testing interface, control software to databasemodule. It finally analyzes Knapp-Kushner detection method of health liquorintelligent visual detection robot, gives test data and applies the technology used tomake health liquor intelligent visual detection robot into transfusion visual detectionrobot, ampoule visual detection robot and oral liquid visual detection robot, whichwere experimented on the spot.
We have solved many a problems in theory and in application during the processof researching on health liquor intelligent visual detection robot, from project designto detection experiment, from image pre-processing and positioning algorithm tovisible foreign matter detection and identification and research on bottle detectionmethod, from machine design to electrical control and transmission, from prototyperesearch to its application. All of these provide valuable experience for theequipment’s following upgrade and development; provide the relative equipmentresearch and development with good reference and guarantee the high quality andsafety of the alcoholic drinks and medicines in the automatic production lines.
本论文旨在研制面向大型高速保健酒自动化生产线上的产品检测智能机器人。保健酒智能视觉检测机器人是以多传感器信息融合为基础，利用机器视觉和智能控制技术，在不影响正常的安全生产过程的同时，可实现高速保健酒自动化生产线上的自动检测，并能把不合格保健酒实时自动分离出生产线。
保健酒智能视觉检测机器人涉及机器视觉、机械设计与传动、光机电一体化、传感器检测与信息融合、光学成像、数字图像处理和通信等多学科领域技术。本文从饮料酒检测现状出发，首先分析了机器人实施的技术路线和整体设计方案，研究了机械结构、机电控制系统、光学成像与视觉信息获取，详细设计开发了针对保健酒液中的可见异物检测和外观检测算法，并经过了大量实验论证，最后研制出保健酒智能视觉检测机器人样机，并把相关技术推广到大输液和口服液等医药自动化生产线上，为饮料医药行业的产品安全生产提供重要保障。归纳起来，本文主要工作和取得的创新成果有：
1、阐述了饮料酒产品的检测现状，指出研制智能视觉检测机器人的重要意义，概述了机器视觉技术，分析了保健酒中可见异物及外观缺陷的成因，重点分析了国内外饮料酒检测的现状。
2、对保健酒智能视觉检测机器人系统进行了可行性分析论证，给出了整体实施技术路线和技术方案，设计了直线间歇式、圆盘固定位置拍照式和旋转跟踪式三种保健酒智能视觉检测机器人的机械结构，并进行了对比分析，研究了基于多IPC和PLC的分布式精密光机电控制结构，分析了伺服跟踪拍摄控制的“加速跟上-同角速度运转-快速返回”三个阶段过程，提出了黑色异物和装量检测工位、白色异物检测工位和瓶盖检测工位的光学成像方法，给出了机器人的检测流程，对比研究了多种次品分拣方案，分析了机器人的软件构架。
3、对保健酒中的可见异物对象进行了运动学分析，并把可见异物和气泡、瓶壁刮痕刻度等进行了对比，对基本滤波算法进行了分析研究了一种改进型加权滤波处理算法，并提出了基于中值的加权均值滤波算法，既能有效的滤除噪声，又能有效的保留运动目标，根据保健酒特定的瓶型进行了检测区域的标定，提出了一种基于边缘统计的经验规则定位法进行图像定位，并得到了良好的定位效果。
4、研究了保健酒液序列图像中可见异物的分割与识别方法。分析了背景减除的运动目标提取方法，详细介绍了基于时间差分的运行目标提取，研究了基于Fisher评价函数的阈值分割，详细分析了基于改进型Mean Shift和Kalman滤波的运动目标跟踪方法，并进行了实例分析与测试，从运动目标的位置、形状和运动轨迹等方面，构建支持向量，研究了基于支持向量机的运动目标识别，进行了图像标定，然后从压像素角度分析，对可见异物进行测量，并判断异物尺寸是否在允许范围内，通过实验，验证了可见异物分割与识别算法的可行性和有效性。
5、研究了保健酒产品的外观检测算法。针对侧瓶盖缺陷，提出了多重扫描的边缘特殊点重心法定位，针对顶面瓶盖，提出了一种多重搜索定位法定位，得到定位点后，研究了一种基于分层思想的瓶盖缺陷图像快速匹配检测方法实现瓶盖检测，针对Sobel算子边缘检测的优点和缺陷，提出了基于改进型Sobel算子，得到液面线，并根据经验规则判断液面高度，实现装量的检测。
6、实现了保健酒智能视觉检测机器人的样机研制及推广应用。给出了机器人的实物图和各关键部分装置图，从人机界面、测试界面、控制软件和数据库模块四个方面分析了软件部分的构成，最后分析了保健酒智能视觉检测机器人的Knapp-Kushner测试方法，给出了测试数据，最后把保健酒智能视觉检测机器人的相关技术推广到大输液视觉检测机器人、安瓿视觉检测机器人和口服液视觉检测机器人等设备上，并得到了实际应用。
在保健酒智能视觉检测机器人研发过程中，从方案设计到测试实验，从图像预处理与定位算法，到可见异物检测与识别和外观检测方法的研究，从机械设计到精密光机电控制与传动，从样机研制到推广应用，发现并解决了大量的理论和实际问题，为设备的应用及后一步的升级开发提供了宝贵的经验，给相关设备的研发带来了良好的参考价值，为实现饮料酒和医药自动化生产线上的产品质量安全提供了保障。
China has a large alcoholic drinks market and their production increased quicklyyear by year, however the automation level of the equipment, especially theautomatic detection, in the alcoholic drinks production line needs improving. Inrecent years, with the increase of consumption and perfection of laws and regulations,the product quality is more and more important in the alcoholic drinks industry. Thenaked-eye detection, which is not efficient and accurate enough, is being challengedby the advanced alcoholic drinks production equipments from foreign countries.Therefore, it is significant to research on alcoholic detection equipment withproprietary intellectual property right in China.
The project is designed to develop an intelligent robot to detect the products inthe large-scale and super-speed health liquor automatic production line. Based onmulti-sensor information fusion detection and recognition, the intelligent robot candetect the products automatically and separate the unqualified products from theproduction line. Meanwhile, the intelligent robot can be applied in other fields, suchas drinks and medicine industries, and insure the safety of products in such fields.
Health liquor intelligent visual detection robot involves technologies in manydisciplines, such as machine vision, machine design and transmission, electricalcontrol, sensor detection and information integration, optical imaging, digital imageprocessing, communication and so on. This paper, firstly, analyzes the overall designproposal and the technologies that the intelligent robot involves. It secondly presentsthe proved algorithm aimed at detection of visible foreign matter in the health liquorand the bottle detection by researching on the mechanical structure, electrical controlsystem, optical imaging, and visual information acquisition. Finally, this paperpresents a prototype of the health liquor intelligent visual detection robot and appliesthe technology into the production of some medicine, such as transfusion and oralliquid. All in all, the innovative achievements presented in this paper are as follows.
1. This paper elaborates the present situation of alcoholic drinks detection,points out the significance of developing an intelligent visual detection robot,introduces machine vision technology, analyzes the reasons why the visible foreignmatter in the health liquor and the bottle defect occur, and focuses on the presentsituation of liquor in China and foreign countries.
2. This paper proves the feasibility of the health liquor intelligent visualdetection robot system, presents the overall technical proposal, compares three kindsof mechanical structure of health liquor intelligent visual detection robot: linearintermittence, taking pictures in a fixed position and rotated tracking, researchesdistributed electrical control structure based on IPC and PLC, analyzes three stagesphotographed by tracking: acceleration, revolve at the same angular speed, rapidreturn, designs the optical imaging that involves detection posts of black foreignmatter, capacity, white foreign matter and bottle cap, presents the robot’s detectionprocedures, compares many kinds of unqualified products sorting methods andanalyzes robot’s software framework.
3. This paper analyzes kinematically the visible foreign matter in the healthliquor, compares visible foreign matter, air bubbles, scratches on the bottle walls, putforwards an improved weighted filter algorithm on the basis of analysis of basicfiltering algorithm, promotes an improved adaptive-filtering algorithm based on theextremum and mean value, which can not only filter efficiently the noise, but alsoremain the moving targets, by researching on the weighted average filteringalgorithm based on mid-value, proposes empirical rule positioning method to positionimages based on edge statistics, which achieves good positioning effect.
4. This paper researches on method of segmentation and identification of visibleforeign matter in the sequential pictures of health liquor, analyzes method ofextraction of moving target with background subtraction, researches on thresholdsegmentation based on Fisher evaluation function, analyzes moving target trackingmethod based on improved Mean Shift and Kalman filtering. This paper also presentsthe living examples to analyze and test. Based on the shape of the moving target,such as length, width, rectangular degree and degree of circularity, as well asestimation of movement, this paper presents support vectors, researches on movingtarget identification on the basis of support vectors, conducts image calibration. Thenanalyzing from the pressure pixel, this paper presents the measurement of visibleforeign matter and the judgment whether the foreign matter size is within the allowedlimits. Finally, this paper presents the feasibility of visible foreign mattersegmentation and identification algorithm through experiments.
5. This paper researches on liquor bottle detection algorithm. Aimed at thedefect that the bottle cape inclines at one side, this paper proposes positioning in theedge special point based on center of gravity and multiple scanning. Directing atdefect that the bottle cap may emboss, this paper promotes a positioning method based on multiple searching. It also researches on detection method of bottle capdefect pictures quickly matching based on layering once the positioning point iscertain. Aimed at the advantages and disadvantages of Sobel operator edge detection,this paper presents an improved Sobel operator to acquire the liquid capacity line andrealize the capacity detection on the basis of liquid surface height judged byexperience.
6. This paper presents the prototype of health liquor intelligent visual detectionrobot and promotes its application. It shows the picture of the robot and set-updiagram of the key components, analyzes the software constitution from four aspectsthat is from human-computer interface, testing interface, control software to databasemodule. It finally analyzes Knapp-Kushner detection method of health liquorintelligent visual detection robot, gives test data and applies the technology used tomake health liquor intelligent visual detection robot into transfusion visual detectionrobot, ampoule visual detection robot and oral liquid visual detection robot, whichwere experimented on the spot.
We have solved many a problems in theory and in application during the processof researching on health liquor intelligent visual detection robot, from project designto detection experiment, from image pre-processing and positioning algorithm tovisible foreign matter detection and identification and research on bottle detectionmethod, from machine design to electrical control and transmission, from prototyperesearch to its application. All of these provide valuable experience for theequipment’s following upgrade and development; provide the relative equipmentresearch and development with good reference and guarantee the high quality andsafety of the alcoholic drinks and medicines in the automatic production lines.
引文
[1]黄书声,韩娜,佟小芳.中国保健酒的历史、现状和发展.酿酒,2008,35(4):16-21
[2]徐发.我国白酒行业的现状及发展趋势.[合肥工业大学硕士学位论文].合肥:合肥工业大学管理学院,2010,5-15
[3]中华人民共和国国家标准. NY/T1508-2007,绿色食品果酒.北京:国家技术监督局,2007
[4]中华人民共和国国家标准. GB/T17204-2008,饮料酒分类.北京:国家技术监督局,2008
[5]郑佳,郭建军,罗惠波.保健酒发展及现状思考.四川食品与发酵,2006,42(3):36-31
[6]沈坤.保健酒市场是方的还是圆的.中国酒,2008,(1):60
[7]沈发治.保健酒浑浊沉淀原因及处理对策.酿酒科技,2009,(12):51-57
[8]吴正奇,凌秀菊.配制型保健酒浑浊沉淀的防止.食品工业,2001,(4):25-26
[9]张烈华,万怀志.保健酒生产工艺探讨.酿酒科技,2006,(9):63-64
[10]刘炯光.保健酒生产中应注意的几个问题.酿酒科技,2005,(6):125-126
[11]中华人民共和国国家标准.GB/T5009.49-2008,发酵酒及其配制酒卫生标准的分析方法.北京:国家技术监督局,2008
[12]李如海.关于民族药酒的沉淀及其澄明度问题的探讨.中国民族医药杂志,2007,13(11):49-50
[13]张玉红,肖文龙.市售白酒抽检情况分析.黑河科技,2008.(1):41-42.
[14]李志斌.白酒中晶状沉淀的研究.酿酒,2008.35(5):22-23.
[15]国家药典委员会.中华人民共和国药典(二部).北京:中国医药科技出版社,2010
[16]段峰,王耀南,雷晓峰,等.机器视觉技术及其应用综述.自动化博览,2002,19(3):59-61
[17]王耀南,李树涛,毛建旭.计算机图像处理与识别技术(第一版).北京:高等教育出版社,2001:56-89.
[18]葛云涛.机器视觉系统集成技术.应用光学,2007,28(2):15-17
[19] Kopparapu S K. Lighting design for machine vision application. Image andVision Computing.2006,24(7):720-726.
[20]李俊.机器视觉照明光源关键技术研究.[硕士学位论文].天津：天津理工大学,2006:22-31
[21] CCS. The Benefits of LED Light Sources. www.machinevisiononline.org/public/articles,2009-02-05
[22] H Grindinger,H Neusser,N Seidenader,V.Wedershoven. Product TestingApparatus. US Patent.0117149,2005-07-02
[23] Anthony James Cronshaw, Christopher James Hodges, Mark RobsonHumphries, etal. Method and Apparatus for Detecting Glass Particles in GlassBottles Filled with Beer. U.S．Patent.6275603B1,2001-8-14
[24]刘焕军.灌装自动化生产线上视觉检测机器人研究:[湖南大学博士学位论文].长沙:湖南大学电气与信息工程学院,2007,4-5,34-41
[25] Brevetti’s Automatic Inspection Machine. http://www.brevetti-cea.com,2011-01-05
[26] Seidenader’s Automatic Inspection Machine. http://www.seidenader.de,2011-01-05
[27] GF’s Automatic Inspection. http://www.gf-industries.it,2011-01-05
[28] Heuft’s products. http://www.heuft.com,2011-06-21
[29] Miho’s products. http://www.miho.de,2011-06-21
[30] Bosch’s Automatic Inspection Machine Products. http://www.bosch.com,2011-01-05
[31] Pharmamech’s Inspection. http://www.pharmamech.com,2011-01-05
[32] Sacmi’s products. http://www.sacmi.it,2011-06-21
[33] Krones’s products. http://www.krones.com,2011-06-21
[34] Eisai’s Inspection Machine. http://www.eisai.com/index.asp,2011-01-05
[35]曲丹丹,罗诗金,薛剑英,等.光阻法智能微粒检测仪的设计与研究.仪器仪表学报,2004,24(4):156-158
[36]张耀,王耀南,周博文.异型瓶药液中可见异物的智能视觉检测机器人.仪器仪表学报,2010,31(5):1058-1063
[37]王耀南,葛继,周博文等.生产线异形瓶装液体中的异物机器视觉识别方法及装置.中国专利. ZL200810143630.9,2011-03-23
[38]程军,崔继波,苟凯英.车辆控制系统CAN总线通信的实施方法.汽车工程,2001,23(5):300-305
[39]敖荣庆,袁坤.伺服系统.北京:航空航天出版社,2007,178-195
[40]张莉松,胡祐德,徐立新.伺服系统原理与设计.第3版.北京:北京理工大学出版社,2008:123-129
[41]郑建伟,张玉,吴秀峰.伦茨交流伺服系统在钢板横剪中位置全闭环的应用.锻压装备与制造技术,2009,44(5):66-68
[42] Feng Duan, Yaonan Wang, Huanjun Liu. A Real-time Machine Vision Systemfor Bottle Finish Inspection. In: Proceedings of Eighth InternationalConference on Control, Automation, Robotics and Vision. Kunming:2004,842-846
[43] Akira Ishii，Takayuki Mizuta，Shigehiko Todo．Detection of Foreign SubstancesMixed in A Plastic Bottle of Medicinal Solution Using Real-time Video ImageProcessing. In: The14th Pattern Recognition Fourteenth InternationalConference. Australia,1998,1646-1650
[44]李杨果,王耀南,王威.基于机器视觉的大输液智能灯检机研究.光电工程,2006.33(11):69-74.
[45]王威.视觉检测系统及其在葡萄糖药液检测中的应用研究.[湖南大学硕士学位论文].长沙：湖南大学电气与信息工程学院,2007:22-38
[46]鲁娟.大输液中可见异物智能检测技术研究.[硕士学位论文].湖南长沙:湖南大学电气与信息工程学院,2008:49-50
[47]杨福刚,孙同景,宋松林.基于机器视觉的全自动灯检机关键技术研究.仪器仪表学报,2008,29(3):562-566.
[48]葛继,王耀南,张辉,等.基于改进型PCNN的智能灯检机研究.仪器仪表学报,2009,30(9):1866-1873
[49]阮秋琦.数字图像处理学.北京:电子工业出版社,2003:4-7,49-50,180-181.
[50] Li Wen-qiang, Ma Fu-chang, Zhang Ying-mei, Yang Lu. A algorithm of noisereduction for ultrasonic image on time difference method based on animproved median filter. In: Intelligent Computing and Intelligent Systems(ICIS),2010IEEE International Conference on2010,861-864
[51] Dali Chen, Dingyu Xue, Feng Pan. An Improved Median Filter Based onAutomatic Parameter Tuning Approach. Mechatronics and Automation,2007.ICMA2007. International Conference on.2007:1305-1309
[52] Deivalakshmi, S, Sarath, S, Palanisamy, P. Detection and removal of Salt andPepper noise in images by improved median filter. In: Recent Advances inIntelligent Computational Systems (RAICS),2011,363-368
[53] Changhong Wang, Taoyi Chen, Zhenshen Qu. A novel improved median filterfor salt-and-pepper noise from highly corrupted images. IN: Systems andControl in Aeronautics and Astronautics (ISSCAA), International Symposiumon2010,3,718-722
[54]刑藏菊,王守觉,邓浩江,等.一种基于极值中值的新型滤波算法.中国图象图形学报,2001,6(6):533-536
[55]窦丽华,毕超.一种快速的图像边缘精确提取算法.光学技术,2006,32(4):496-499
[56] Feng Duan, Yaonan Wang, Wei Duan. Super Dynamic CCD Camera Based OnMulti-sensor Image Fusion. In: Proceedings of the4th World Congress onIntelligent Control and Automation(6). Shanghai:2002,2211-2219
[57]高颖慧,李吉成,沈振康.海空背景下红外运动小目标的检测方法.系统工程与电子技术,2004,26(6):741-743
[58]沈宇键,何昕,郝志航.国象序列中检测运动小目标的递归算法.光电工程,2000,27(2):9-13
[59]过润秋,李大鹏,林晓春.红外点目标检测的小波变换方法研究.光子学报,2004,33(4):464-467
[60] Xu Jiping, Ikram-ul-Haq, Chen Jie, Dou Lihua, Liu Zaiwen. Moving TargetDetection and Tracking in FLIR Image Sequences Based on Thermal TargetModeling. In: Measuring Technology and Mechatronics Automation(ICMTMA),2010International Conference on.2010,715-720
[61]徐永兵,裴先登.红外序列图像中运动小目标的检测.华中科技大学学报,2004,32(1):67-69
[62]成玉娟.液体中小目标检测算法研究及应用.[浙江大学硕士学位论文].浙江杭州:浙江大学电气与信息工程学院,2002
[63]聂烜,陈怀民,朱怡安.一种用于目标检测的复合差分法.西北工业大学学报,2009,27(1):88-9
[64]张文超,王岩飞,陈贺新.基于Tophat变换的复杂背景下运动点的目标识别算法.中国图象图形学报,2007,12(5):871-874
[65] H．Yamazaki, T．Katane，H．Fukuda，K．Asano，M．Matsushima．InspectionDevice and System for Inspecting Foreign Matters in A Liquid FilledTransparent Container.US Patent.6937339,2005-08-30
[66] Robert T Collins, Alan J Lipton, Takeo Kanade. A system for videoSurveillance and Monitoring. Carnegie Mellon:Carnegie Mellon University,2000:54-67
[67] Paragios N, Deriche R. Geodesic active contours and level sets for thedetection andtracking of moving objects. IEEE Transactions on PatternAnalysis and Machine Intelligence.2000,22(3):266-280
[68]韩思奇,王蕾.图像分割的阈值法综述.系统工程与电子技术,2002,24(6):91-94.
[69] Xu Zhiqian, Yan Xiangzhen, Yang Xiujuan. A new recognition andinterpretation system of the blurred image. In: Electric Technology and CivilEngineering (ICETCE),2011International Conference on.2011,958-961
[70]鲜海滢,李晓峰,李在铭.基于区域分割的噪声抑制及红外目标检测.红外与毫米波学报,2008,27(4):269-274
[71]贾振堂,贺贵明,韩艳芳.运动视频对象分割的一种快速算法.中国中国图象图形学报,2002,7(11):1123-1127
[72] Sezgin M, BülentSankur. Survey over image thresholding techniques andquantitative performance evaluation. Journal of Electronic Imaging.2004,13(1):146-168.
[73] Rais N B, Hanif M S, Taj I A. Adaptive thresholding technique for documentimage analysis. In: Multitopic Conference Proceedings of INMIC20048thInternational,2004,61-66
[74] Zhao YL, Zhang ZC, Gao ZM, A simple and workable moving objectssegmentation method. In: Electronics in Marine,2004.2004,585-590.
[75] Ming Hong Pi, Hong Zhang. Two-stage image segmentation by adaptivethresholding and gradient watershed. Computer and Robot Vision Proceedings,2005,57-64
[76] James Fung,Steve Mann. Computer vision signal processing on graphicsprocessing units.2004,93-96
[77]张雄美,李钊,易昭湘,等.基于局部模糊熵与Otsu的图像阈值分割.无线电通信技术,2007,33(5):38-40
[78]陈琪,熊博莅,陆军等.二维类内最小交叉熵的图像分割快速方法.计算机工程与应用,2011,47(9):149-151,203
[79]陈果.图像阈值分割的Fisher准则函数法.仪器仪表学报,2003.24(6):564-568
[80]童莹,邱晓晖.基于Fisher准则函数的二维阈值图像分割算法.电力系统通信,2004.9(9):36-40.
[81] Wei Huang, Liqin Fu, Yu Xiao. Shadow removal based on invariant image withFisher discrimination criterion. Image Analysis and Signal Processing (IASP),2011International Conference on.2011,214-218
[82]刘钢,刘明,匡海鹏.多目标跟踪方法综述.电光与控制,2004,11(3):26-29
[83] LiP H, Zhang T W and Pece A. Visual contour tracking based on particle filters.Image and Vision Computing,2003,21(1):111-123
[84]万琴.智能视觉监控中多运动目标检测与跟踪方法研究.[湖南大学博士学位论文].长沙：湖南大学电气与信息工程学院,2009,67-72
[85] DeCarlo D and Metaxas D. Optical flow constraints on deformable models withapplications to face tracking. International Journal of Computer Vision,2000,38(2):99-127
[86] Schoepflin, T., Chalana, V., Haynor, D.R., Yongmin Kim, Video objecttracking with a sequential hierarchy of template deformations. In: Circuits andSystems for Video Technology, IEEE Transactions on.2001,11(11):1171-1182
[87]宋新,王鲁平,王平,等.基于改进均值位移的红外目标跟踪方法.红外与毫米波学报,2007,26(6):429-432
[88]蔡征,黄瑞光.运动图像序列的多目标跟踪技术及实现.计算机与数字工程,2006,34(9):140-143
[89]彭宁嵩,杨杰,刘志,等. Mean-Shift跟踪算法中核函数窗宽的自动选取.软件学报,2005,16(9):1542-1550
[90]颜佳,吴敏渊,陈淑珍,等.应用Mean Shift和分块的抗遮挡跟踪.光学精密工程,2010,18(6):1413-1419
[91]赵倩,袁健全,鲁新平.结合目标预估计与Mean Shift理论的运动目标跟踪算法.红外与激光工程,2010,39(6):1152-1156
[92]李培华.一种改进的Mean Shift跟踪算法.自动化学报,2007,33(4):347-354
[93]薛陈,朱明,陈爱华.鲁棒的基于改进Mean-shift的目标跟踪.光学精密工程,2010,18(1):234-238
[94]文志强,蔡自兴. Mean Shift算法的收敛性分析.软件学报,2007,18(2):1112-1118
[95] Jamshaid A, Muhammad U. A Consistent and Robust Kalman Filter Design forIn-motion Alignment of Inertial Navigation System. Measurement,2009,42(4):577-582
[96] Mohinder S. Grewal,Angus P. Andrew.Kalman Filtering Theory and Practiceusing MATLAB(3rd Edition).New York:Wiley-IEEE Press,2008,1-5
[97]万琴,王耀南.基于卡尔曼滤波器的运动目标检测与跟踪.湖南大学学报(自然科学版),2007,34(3):36-40
[98]虞旦,韦巍,张远辉.一种基于卡尔曼预测的动态目标跟踪算法研究.光电工程,2009,36(1):52-56
[99]吕娜,冯祖仁.质心迭代图像跟踪算法.西安交通大学学报,2007,41(2):1396-1400
[100] Zhiwei Liang，Xudong Ma，Xianzhong Dai．Robust tracking of Moving SoundSource Using Multiple Model Kalman Filter．Applied Acoustics，2008，69(12):1350-1355
[101] Mohinder S. Grewal,Angus P. Andrew.Kalman Filtering Theory and Practiceusing MATLAB(3rd Edition).New York:Wiley-IEEE Press,2008,1-5
[102] Jamshaid Ali，Muhammad Ushaq．A Consistent and Robust Kalman FilterDesign for In-motion Alignment of Inertial Navigation System．Measurement，2009,42(4):577-582
[103] Huanjun Liu, Yaonan Wang, Feng Duan. An Empty Bottle Intelligent InspectorBased on Support Vector Machines and Fuzzy Theory. In: ConferenceProceedings of Sixth World Congress on Intellignent Control and Automation.2006,9739-9743.
[104] Yaonan Wang, Huanjun Liu, Feng Duan. A Bottle Finish Inspect Method Basedon Fuzzy Support Vector Machines and Wavelet Transform. IN: Proceedings of2005International Conference on Machine Learning and Cybernetics2005,4588-4592
[105] Huanjun Liu, Yaonan Wang. A Method to Choose Kernel Function and ItsParameters for Support Vector Machine. Proceedings of2005InternationalConference on Machine Learning and Cybernetics2005:4277-4280.
[106] Dequn Zhao, Weiwei Zou, GuangMin Sun. A fast image classificationalgorithm using Support Vector Machine. In: Computer Technology andDevelopment (ICCTD), International Conference on.2010,2:385-388
[107] Lan Gan, Zhongping Yu. Based on Support Vector Machine's Tumor ImageClassifier Design. In:e-Education, e-Business, e-Management, and e-Learning,2010. IC4E '10. International Conference on.2010,137-140
[108] Jia Xiaofen, Zhao Baiting, Chen Zhaoquan. Genetic algorithm optimizationbased on support vector machine image interpolation. In: Cross StraitQuad-Regional Radio Science and Wireless Technology Conference(CSQRWC),2011,1319-1322
[109] Lan Gan, Zhongping Yu, Jie Gao. Global Feature-Based Image Classificationand Recognition in Small Sample Size Problem. In: e-Education, e-Business,e-Management, and e-Learning,2010. IC4E '10. International Conference on.2010,30-33
[110] Habib, T, Inglada, J, Mercier, G, Chanussot, J. Speeding up Support VectorMachine (SVM) image classification by a kernel series expansion. In: ImageProcessing. IEEE International Conference on.2008,15:865-868
[111] Zhao Baoyong, Qi Yingjian. Image classification with ant colony based supportvector machine. In: Control Conference (CCC).2011,30:3260–3263.
[112] Hu Z, Cai Y, Li Y, et al. Support vector machine based ensemble classifier.In:The2005American Control Conference.2005,745-749
[113] Smith R S, Kittler J, Hamouz M, et al. Face Recognition Using Angular LDAand SVM Ensembles.In: The18th International Conference on PatternRecognition.2006,1008-1012
[114]邾继贵,叶声华.基于近景数字摄影的坐标精密测量关键技术研究.计量学报,2005,26(3):207-211
[115]章毓晋.图像处理和分析基础.北京:高等教育出版社,2002:180-181
[116] Rafael C. Gonzalez,Richard E. Woods.数字图像处理(第二版).北京:电子工业出版社,2003:496-500,215-219
[117] Liangyu Lei. A Machine Vision System for Inspecting Bearing-diameter. In:Proceedings of the5th World Congress on Intelligent Control and Automation.Hangzhou:2004,3904-3906
[118] Cheng YJ, Yan HM, Hui H. A novel algorithm for pixel-target detection inliquid image. Acta Photonica Sinica2002,31(6):743-747
[119]段峰,王耀南,刘焕军.基于机器视觉的智能空瓶检测机器人研究.仪器仪表学报,2004,25(5):624-627.
[120] Bartleet Trevor C., de Jager Gerhard. Real-time machine vision inspection ofplastic bottle closures. Proceedings of the SPIE,.1995,2598:362-372
[121] Hui-min Ma, Guang-da Su, Jun-yan Wang et al. A Glass Bottle DefectDetection System Without Touching. In: Proceedings of the first InternationalConference on Machine Learning and Cybernetics, Beijing, China,2002,628-632
[122] Kumar A., ang G.K.H. Defect detection in textured materials using optimizedfilters. IEEE Transactions on Systems, Man and Cybernetics, Part B,2002,2(5):553–570
[123] Basu M. Gaussian-based edge-detection methods-a survey. IEEE Transactionson Systems, Man and Cybernetics.2002,32(3):252-260
[124]范俐捷,王岩飞,高鑫.一种新的基于灰度的图像匹配方法.微计算机信息,2007,30(8):296-297
[125]杨小冈,缪栋,付光远.一种基于图象物理特征的分层匹配算法.计算机工程与应用,2002,15(10):106-108
[126] Sakurai K., Onoyama A., Fujii T., et al. Solution of pattern matchinginspection problem for grainy metal layers. IEEE Transactions onSemiconductor Manufacturing,2002,15(1):118–126
[127]靳鹏飞.一种改进的Sobel图像边缘检测算法.应用光学,2008，29(4):625-628
[128]姜炳旭,刘杰,孙可. Sobel边缘检测的细化.沈阳师范大学学报(自然科学版),2010,28(4):503-506
[129] Caixia Deng, Weifeng Ma, Yin Yin. An edge detection approach of imagefusion based on improved Sobel operator. In: Image and Signal Processing(CISP), International Congress on.2011,4:1189–1193
[130] Chunxi Ma; Lei Yang; Wenshuo Gao; Zhonghui Liu. An improved Sobelalgorithm based on median filter. In: Mechanical and Electronics Engineering(ICMEE), International Conference on.2010,2: V1-88-V1-92
[131] Wenshuo Gao, Xiaoguang Zhang, Lei Yang, Huizhong Liu. An improved Sobeledge detection. In: Computer Science and Information Technology (ICCSIT),IEEE International Conference on.2010,3:67-71
[132]段峰.啤酒瓶视觉检测机器人研究.[湖南大学博士学位论文].长沙:湖南大学电气与信息工程学院,2007,88-96
[2]徐发.我国白酒行业的现状及发展趋势.[合肥工业大学硕士学位论文].合肥:合肥工业大学管理学院,2010,5-15
[3]中华人民共和国国家标准. NY/T1508-2007,绿色食品果酒.北京:国家技术监督局,2007
[4]中华人民共和国国家标准. GB/T17204-2008,饮料酒分类.北京:国家技术监督局,2008
[5]郑佳,郭建军,罗惠波.保健酒发展及现状思考.四川食品与发酵,2006,42(3):36-31
[6]沈坤.保健酒市场是方的还是圆的.中国酒,2008,(1):60
[7]沈发治.保健酒浑浊沉淀原因及处理对策.酿酒科技,2009,(12):51-57
[8]吴正奇,凌秀菊.配制型保健酒浑浊沉淀的防止.食品工业,2001,(4):25-26
[9]张烈华,万怀志.保健酒生产工艺探讨.酿酒科技,2006,(9):63-64
[10]刘炯光.保健酒生产中应注意的几个问题.酿酒科技,2005,(6):125-126
[11]中华人民共和国国家标准.GB/T5009.49-2008,发酵酒及其配制酒卫生标准的分析方法.北京:国家技术监督局,2008
[12]李如海.关于民族药酒的沉淀及其澄明度问题的探讨.中国民族医药杂志,2007,13(11):49-50
[13]张玉红,肖文龙.市售白酒抽检情况分析.黑河科技,2008.(1):41-42.
[14]李志斌.白酒中晶状沉淀的研究.酿酒,2008.35(5):22-23.
[15]国家药典委员会.中华人民共和国药典(二部).北京:中国医药科技出版社,2010
[16]段峰,王耀南,雷晓峰,等.机器视觉技术及其应用综述.自动化博览,2002,19(3):59-61
[17]王耀南,李树涛,毛建旭.计算机图像处理与识别技术(第一版).北京:高等教育出版社,2001:56-89.
[18]葛云涛.机器视觉系统集成技术.应用光学,2007,28(2):15-17
[19] Kopparapu S K. Lighting design for machine vision application. Image andVision Computing.2006,24(7):720-726.
[20]李俊.机器视觉照明光源关键技术研究.[硕士学位论文].天津：天津理工大学,2006:22-31
[21] CCS. The Benefits of LED Light Sources. www.machinevisiononline.org/public/articles,2009-02-05
[22] H Grindinger,H Neusser,N Seidenader,V.Wedershoven. Product TestingApparatus. US Patent.0117149,2005-07-02
[23] Anthony James Cronshaw, Christopher James Hodges, Mark RobsonHumphries, etal. Method and Apparatus for Detecting Glass Particles in GlassBottles Filled with Beer. U.S．Patent.6275603B1,2001-8-14
[24]刘焕军.灌装自动化生产线上视觉检测机器人研究:[湖南大学博士学位论文].长沙:湖南大学电气与信息工程学院,2007,4-5,34-41
[25] Brevetti’s Automatic Inspection Machine. http://www.brevetti-cea.com,2011-01-05
[26] Seidenader’s Automatic Inspection Machine. http://www.seidenader.de,2011-01-05
[27] GF’s Automatic Inspection. http://www.gf-industries.it,2011-01-05
[28] Heuft’s products. http://www.heuft.com,2011-06-21
[29] Miho’s products. http://www.miho.de,2011-06-21
[30] Bosch’s Automatic Inspection Machine Products. http://www.bosch.com,2011-01-05
[31] Pharmamech’s Inspection. http://www.pharmamech.com,2011-01-05
[32] Sacmi’s products. http://www.sacmi.it,2011-06-21
[33] Krones’s products. http://www.krones.com,2011-06-21
[34] Eisai’s Inspection Machine. http://www.eisai.com/index.asp,2011-01-05
[35]曲丹丹,罗诗金,薛剑英,等.光阻法智能微粒检测仪的设计与研究.仪器仪表学报,2004,24(4):156-158
[36]张耀,王耀南,周博文.异型瓶药液中可见异物的智能视觉检测机器人.仪器仪表学报,2010,31(5):1058-1063
[37]王耀南,葛继,周博文等.生产线异形瓶装液体中的异物机器视觉识别方法及装置.中国专利. ZL200810143630.9,2011-03-23
[38]程军,崔继波,苟凯英.车辆控制系统CAN总线通信的实施方法.汽车工程,2001,23(5):300-305
[39]敖荣庆,袁坤.伺服系统.北京:航空航天出版社,2007,178-195
[40]张莉松,胡祐德,徐立新.伺服系统原理与设计.第3版.北京:北京理工大学出版社,2008:123-129
[41]郑建伟,张玉,吴秀峰.伦茨交流伺服系统在钢板横剪中位置全闭环的应用.锻压装备与制造技术,2009,44(5):66-68
[42] Feng Duan, Yaonan Wang, Huanjun Liu. A Real-time Machine Vision Systemfor Bottle Finish Inspection. In: Proceedings of Eighth InternationalConference on Control, Automation, Robotics and Vision. Kunming:2004,842-846
[43] Akira Ishii，Takayuki Mizuta，Shigehiko Todo．Detection of Foreign SubstancesMixed in A Plastic Bottle of Medicinal Solution Using Real-time Video ImageProcessing. In: The14th Pattern Recognition Fourteenth InternationalConference. Australia,1998,1646-1650
[44]李杨果,王耀南,王威.基于机器视觉的大输液智能灯检机研究.光电工程,2006.33(11):69-74.
[45]王威.视觉检测系统及其在葡萄糖药液检测中的应用研究.[湖南大学硕士学位论文].长沙：湖南大学电气与信息工程学院,2007:22-38
[46]鲁娟.大输液中可见异物智能检测技术研究.[硕士学位论文].湖南长沙:湖南大学电气与信息工程学院,2008:49-50
[47]杨福刚,孙同景,宋松林.基于机器视觉的全自动灯检机关键技术研究.仪器仪表学报,2008,29(3):562-566.
[48]葛继,王耀南,张辉,等.基于改进型PCNN的智能灯检机研究.仪器仪表学报,2009,30(9):1866-1873
[49]阮秋琦.数字图像处理学.北京:电子工业出版社,2003:4-7,49-50,180-181.
[50] Li Wen-qiang, Ma Fu-chang, Zhang Ying-mei, Yang Lu. A algorithm of noisereduction for ultrasonic image on time difference method based on animproved median filter. In: Intelligent Computing and Intelligent Systems(ICIS),2010IEEE International Conference on2010,861-864
[51] Dali Chen, Dingyu Xue, Feng Pan. An Improved Median Filter Based onAutomatic Parameter Tuning Approach. Mechatronics and Automation,2007.ICMA2007. International Conference on.2007:1305-1309
[52] Deivalakshmi, S, Sarath, S, Palanisamy, P. Detection and removal of Salt andPepper noise in images by improved median filter. In: Recent Advances inIntelligent Computational Systems (RAICS),2011,363-368
[53] Changhong Wang, Taoyi Chen, Zhenshen Qu. A novel improved median filterfor salt-and-pepper noise from highly corrupted images. IN: Systems andControl in Aeronautics and Astronautics (ISSCAA), International Symposiumon2010,3,718-722
[54]刑藏菊,王守觉,邓浩江,等.一种基于极值中值的新型滤波算法.中国图象图形学报,2001,6(6):533-536
[55]窦丽华,毕超.一种快速的图像边缘精确提取算法.光学技术,2006,32(4):496-499
[56] Feng Duan, Yaonan Wang, Wei Duan. Super Dynamic CCD Camera Based OnMulti-sensor Image Fusion. In: Proceedings of the4th World Congress onIntelligent Control and Automation(6). Shanghai:2002,2211-2219
[57]高颖慧,李吉成,沈振康.海空背景下红外运动小目标的检测方法.系统工程与电子技术,2004,26(6):741-743
[58]沈宇键,何昕,郝志航.国象序列中检测运动小目标的递归算法.光电工程,2000,27(2):9-13
[59]过润秋,李大鹏,林晓春.红外点目标检测的小波变换方法研究.光子学报,2004,33(4):464-467
[60] Xu Jiping, Ikram-ul-Haq, Chen Jie, Dou Lihua, Liu Zaiwen. Moving TargetDetection and Tracking in FLIR Image Sequences Based on Thermal TargetModeling. In: Measuring Technology and Mechatronics Automation(ICMTMA),2010International Conference on.2010,715-720
[61]徐永兵,裴先登.红外序列图像中运动小目标的检测.华中科技大学学报,2004,32(1):67-69
[62]成玉娟.液体中小目标检测算法研究及应用.[浙江大学硕士学位论文].浙江杭州:浙江大学电气与信息工程学院,2002
[63]聂烜,陈怀民,朱怡安.一种用于目标检测的复合差分法.西北工业大学学报,2009,27(1):88-9
[64]张文超,王岩飞,陈贺新.基于Tophat变换的复杂背景下运动点的目标识别算法.中国图象图形学报,2007,12(5):871-874
[65] H．Yamazaki, T．Katane，H．Fukuda，K．Asano，M．Matsushima．InspectionDevice and System for Inspecting Foreign Matters in A Liquid FilledTransparent Container.US Patent.6937339,2005-08-30
[66] Robert T Collins, Alan J Lipton, Takeo Kanade. A system for videoSurveillance and Monitoring. Carnegie Mellon:Carnegie Mellon University,2000:54-67
[67] Paragios N, Deriche R. Geodesic active contours and level sets for thedetection andtracking of moving objects. IEEE Transactions on PatternAnalysis and Machine Intelligence.2000,22(3):266-280
[68]韩思奇,王蕾.图像分割的阈值法综述.系统工程与电子技术,2002,24(6):91-94.
[69] Xu Zhiqian, Yan Xiangzhen, Yang Xiujuan. A new recognition andinterpretation system of the blurred image. In: Electric Technology and CivilEngineering (ICETCE),2011International Conference on.2011,958-961
[70]鲜海滢,李晓峰,李在铭.基于区域分割的噪声抑制及红外目标检测.红外与毫米波学报,2008,27(4):269-274
[71]贾振堂,贺贵明,韩艳芳.运动视频对象分割的一种快速算法.中国中国图象图形学报,2002,7(11):1123-1127
[72] Sezgin M, BülentSankur. Survey over image thresholding techniques andquantitative performance evaluation. Journal of Electronic Imaging.2004,13(1):146-168.
[73] Rais N B, Hanif M S, Taj I A. Adaptive thresholding technique for documentimage analysis. In: Multitopic Conference Proceedings of INMIC20048thInternational,2004,61-66
[74] Zhao YL, Zhang ZC, Gao ZM, A simple and workable moving objectssegmentation method. In: Electronics in Marine,2004.2004,585-590.
[75] Ming Hong Pi, Hong Zhang. Two-stage image segmentation by adaptivethresholding and gradient watershed. Computer and Robot Vision Proceedings,2005,57-64
[76] James Fung,Steve Mann. Computer vision signal processing on graphicsprocessing units.2004,93-96
[77]张雄美,李钊,易昭湘,等.基于局部模糊熵与Otsu的图像阈值分割.无线电通信技术,2007,33(5):38-40
[78]陈琪,熊博莅,陆军等.二维类内最小交叉熵的图像分割快速方法.计算机工程与应用,2011,47(9):149-151,203
[79]陈果.图像阈值分割的Fisher准则函数法.仪器仪表学报,2003.24(6):564-568
[80]童莹,邱晓晖.基于Fisher准则函数的二维阈值图像分割算法.电力系统通信,2004.9(9):36-40.
[81] Wei Huang, Liqin Fu, Yu Xiao. Shadow removal based on invariant image withFisher discrimination criterion. Image Analysis and Signal Processing (IASP),2011International Conference on.2011,214-218
[82]刘钢,刘明,匡海鹏.多目标跟踪方法综述.电光与控制,2004,11(3):26-29
[83] LiP H, Zhang T W and Pece A. Visual contour tracking based on particle filters.Image and Vision Computing,2003,21(1):111-123
[84]万琴.智能视觉监控中多运动目标检测与跟踪方法研究.[湖南大学博士学位论文].长沙：湖南大学电气与信息工程学院,2009,67-72
[85] DeCarlo D and Metaxas D. Optical flow constraints on deformable models withapplications to face tracking. International Journal of Computer Vision,2000,38(2):99-127
[86] Schoepflin, T., Chalana, V., Haynor, D.R., Yongmin Kim, Video objecttracking with a sequential hierarchy of template deformations. In: Circuits andSystems for Video Technology, IEEE Transactions on.2001,11(11):1171-1182
[87]宋新,王鲁平,王平,等.基于改进均值位移的红外目标跟踪方法.红外与毫米波学报,2007,26(6):429-432
[88]蔡征,黄瑞光.运动图像序列的多目标跟踪技术及实现.计算机与数字工程,2006,34(9):140-143
[89]彭宁嵩,杨杰,刘志,等. Mean-Shift跟踪算法中核函数窗宽的自动选取.软件学报,2005,16(9):1542-1550
[90]颜佳,吴敏渊,陈淑珍,等.应用Mean Shift和分块的抗遮挡跟踪.光学精密工程,2010,18(6):1413-1419
[91]赵倩,袁健全,鲁新平.结合目标预估计与Mean Shift理论的运动目标跟踪算法.红外与激光工程,2010,39(6):1152-1156
[92]李培华.一种改进的Mean Shift跟踪算法.自动化学报,2007,33(4):347-354
[93]薛陈,朱明,陈爱华.鲁棒的基于改进Mean-shift的目标跟踪.光学精密工程,2010,18(1):234-238
[94]文志强,蔡自兴. Mean Shift算法的收敛性分析.软件学报,2007,18(2):1112-1118
[95] Jamshaid A, Muhammad U. A Consistent and Robust Kalman Filter Design forIn-motion Alignment of Inertial Navigation System. Measurement,2009,42(4):577-582
[96] Mohinder S. Grewal,Angus P. Andrew.Kalman Filtering Theory and Practiceusing MATLAB(3rd Edition).New York:Wiley-IEEE Press,2008,1-5
[97]万琴,王耀南.基于卡尔曼滤波器的运动目标检测与跟踪.湖南大学学报(自然科学版),2007,34(3):36-40
[98]虞旦,韦巍,张远辉.一种基于卡尔曼预测的动态目标跟踪算法研究.光电工程,2009,36(1):52-56
[99]吕娜,冯祖仁.质心迭代图像跟踪算法.西安交通大学学报,2007,41(2):1396-1400
[100] Zhiwei Liang，Xudong Ma，Xianzhong Dai．Robust tracking of Moving SoundSource Using Multiple Model Kalman Filter．Applied Acoustics，2008，69(12):1350-1355
[101] Mohinder S. Grewal,Angus P. Andrew.Kalman Filtering Theory and Practiceusing MATLAB(3rd Edition).New York:Wiley-IEEE Press,2008,1-5
[102] Jamshaid Ali，Muhammad Ushaq．A Consistent and Robust Kalman FilterDesign for In-motion Alignment of Inertial Navigation System．Measurement，2009,42(4):577-582
[103] Huanjun Liu, Yaonan Wang, Feng Duan. An Empty Bottle Intelligent InspectorBased on Support Vector Machines and Fuzzy Theory. In: ConferenceProceedings of Sixth World Congress on Intellignent Control and Automation.2006,9739-9743.
[104] Yaonan Wang, Huanjun Liu, Feng Duan. A Bottle Finish Inspect Method Basedon Fuzzy Support Vector Machines and Wavelet Transform. IN: Proceedings of2005International Conference on Machine Learning and Cybernetics2005,4588-4592
[105] Huanjun Liu, Yaonan Wang. A Method to Choose Kernel Function and ItsParameters for Support Vector Machine. Proceedings of2005InternationalConference on Machine Learning and Cybernetics2005:4277-4280.
[106] Dequn Zhao, Weiwei Zou, GuangMin Sun. A fast image classificationalgorithm using Support Vector Machine. In: Computer Technology andDevelopment (ICCTD), International Conference on.2010,2:385-388
[107] Lan Gan, Zhongping Yu. Based on Support Vector Machine's Tumor ImageClassifier Design. In:e-Education, e-Business, e-Management, and e-Learning,2010. IC4E '10. International Conference on.2010,137-140
[108] Jia Xiaofen, Zhao Baiting, Chen Zhaoquan. Genetic algorithm optimizationbased on support vector machine image interpolation. In: Cross StraitQuad-Regional Radio Science and Wireless Technology Conference(CSQRWC),2011,1319-1322
[109] Lan Gan, Zhongping Yu, Jie Gao. Global Feature-Based Image Classificationand Recognition in Small Sample Size Problem. In: e-Education, e-Business,e-Management, and e-Learning,2010. IC4E '10. International Conference on.2010,30-33
[110] Habib, T, Inglada, J, Mercier, G, Chanussot, J. Speeding up Support VectorMachine (SVM) image classification by a kernel series expansion. In: ImageProcessing. IEEE International Conference on.2008,15:865-868
[111] Zhao Baoyong, Qi Yingjian. Image classification with ant colony based supportvector machine. In: Control Conference (CCC).2011,30:3260–3263.
[112] Hu Z, Cai Y, Li Y, et al. Support vector machine based ensemble classifier.In:The2005American Control Conference.2005,745-749
[113] Smith R S, Kittler J, Hamouz M, et al. Face Recognition Using Angular LDAand SVM Ensembles.In: The18th International Conference on PatternRecognition.2006,1008-1012
[114]邾继贵,叶声华.基于近景数字摄影的坐标精密测量关键技术研究.计量学报,2005,26(3):207-211
[115]章毓晋.图像处理和分析基础.北京:高等教育出版社,2002:180-181
[116] Rafael C. Gonzalez,Richard E. Woods.数字图像处理(第二版).北京:电子工业出版社,2003:496-500,215-219
[117] Liangyu Lei. A Machine Vision System for Inspecting Bearing-diameter. In:Proceedings of the5th World Congress on Intelligent Control and Automation.Hangzhou:2004,3904-3906
[118] Cheng YJ, Yan HM, Hui H. A novel algorithm for pixel-target detection inliquid image. Acta Photonica Sinica2002,31(6):743-747
[119]段峰,王耀南,刘焕军.基于机器视觉的智能空瓶检测机器人研究.仪器仪表学报,2004,25(5):624-627.
[120] Bartleet Trevor C., de Jager Gerhard. Real-time machine vision inspection ofplastic bottle closures. Proceedings of the SPIE,.1995,2598:362-372
[121] Hui-min Ma, Guang-da Su, Jun-yan Wang et al. A Glass Bottle DefectDetection System Without Touching. In: Proceedings of the first InternationalConference on Machine Learning and Cybernetics, Beijing, China,2002,628-632
[122] Kumar A., ang G.K.H. Defect detection in textured materials using optimizedfilters. IEEE Transactions on Systems, Man and Cybernetics, Part B,2002,2(5):553–570
[123] Basu M. Gaussian-based edge-detection methods-a survey. IEEE Transactionson Systems, Man and Cybernetics.2002,32(3):252-260
[124]范俐捷,王岩飞,高鑫.一种新的基于灰度的图像匹配方法.微计算机信息,2007,30(8):296-297
[125]杨小冈,缪栋,付光远.一种基于图象物理特征的分层匹配算法.计算机工程与应用,2002,15(10):106-108
[126] Sakurai K., Onoyama A., Fujii T., et al. Solution of pattern matchinginspection problem for grainy metal layers. IEEE Transactions onSemiconductor Manufacturing,2002,15(1):118–126
[127]靳鹏飞.一种改进的Sobel图像边缘检测算法.应用光学,2008，29(4):625-628
[128]姜炳旭,刘杰,孙可. Sobel边缘检测的细化.沈阳师范大学学报(自然科学版),2010,28(4):503-506
[129] Caixia Deng, Weifeng Ma, Yin Yin. An edge detection approach of imagefusion based on improved Sobel operator. In: Image and Signal Processing(CISP), International Congress on.2011,4:1189–1193
[130] Chunxi Ma; Lei Yang; Wenshuo Gao; Zhonghui Liu. An improved Sobelalgorithm based on median filter. In: Mechanical and Electronics Engineering(ICMEE), International Conference on.2010,2: V1-88-V1-92
[131] Wenshuo Gao, Xiaoguang Zhang, Lei Yang, Huizhong Liu. An improved Sobeledge detection. In: Computer Science and Information Technology (ICCSIT),IEEE International Conference on.2010,3:67-71
[132]段峰.啤酒瓶视觉检测机器人研究.[湖南大学博士学位论文].长沙:湖南大学电气与信息工程学院,2007,88-96