基于AdaBoost的集装箱检测方法研究
|
摘要
目前,采用计算机视觉技术是集装箱自动化装卸作业中实现集装箱识别和定位操作的发展方向。本文分析和探讨了两种集装箱检测方法的基本理论,搭建了集装箱装卸模型系统(Container Handling Model System--CHMS),在Windows环境下编写软件算法,并在CHMS系统上对分别对两种方法进行了仿真实验。实验表明本文提出的集装箱检测方法是合理的,具有一定的理论价值与实用价值。本文的研究工作主要包括以下几个方面:
1.将基于AdaBoost的学习算法用于集装箱检测,从一个较大的特征集中选择少量关键的haar-like特征,产生一个高效的强分类器。再用级联方式将单个的强分类器级联成为一个更加复杂的级联分类器。对AdaBoost算法的收敛性能、泛化能力以及权重更新方法对分类器性能的影响等进行了深入的分析。本文使用了自己创建的集装箱模型样本库进行训练得到分类器,并在多种背景、光线条件下进行了集装箱模型检测实验。实验结果表明,得到的集装箱模型分类器效果理想。
2.提出基于Kalman滤波的彩色模板更新的方法。介绍了基于色彩直方图的模板匹配方法和Kalman滤波的基本原理,给出了色彩直方图的具体算法。对模板内每一个像素值的三个颜色分量使用Kalman滤波进行更新,得到最优的模板图像,然后用模板匹配算法来检测集装箱。此方法与原先的固定模板匹配算法相比,提高了模板匹配的准确度和稳定度。
Currently, using computer vision technology is the direction of container recognition and position in automatic container loading and unloading operations.Two of literatures, surveys and research papers concerning up-to-date techniques of container detection are read and analyzed. Putting up the Container Handling Model System(CHMS), compiling the code in Windows, and carrying out simulations about the two methods on the CHMS. The experiments indicate that the methods of container detection proposed in showing a certain degree of theoretical and practical value. Paper mainly includes the following several aspects:
1.Using container detection method based on AdaBoost learning algorithm, which selects few key haar-like features from a large set of features, to build a robust cascade classifier. The convergence and generalization capability of AdaBoost algorithm as well as the effect of weight-updating approach on the classifier's performance are analysed comprehensively. This paper used the sample storehouse which one founded to carry on the training and get the container classifier. Detecting the container model In a variety of backgrounds and lighting conditions used the classifier. The experimental results show that the effect of container classifier is suitable to detection.
2. Providing an approach to update the template based on a Kalman filter. Intruducing the the basic principles of the template matching based on the color histogram and the Kalman filter, and giving the specific algorithm of the color histogram. The Kalman filter update three component of each pixel value and get the optimum template, and containers are detected by the optimum template. This method improves the accuracy and the stability of template matching.
1.将基于AdaBoost的学习算法用于集装箱检测,从一个较大的特征集中选择少量关键的haar-like特征,产生一个高效的强分类器。再用级联方式将单个的强分类器级联成为一个更加复杂的级联分类器。对AdaBoost算法的收敛性能、泛化能力以及权重更新方法对分类器性能的影响等进行了深入的分析。本文使用了自己创建的集装箱模型样本库进行训练得到分类器,并在多种背景、光线条件下进行了集装箱模型检测实验。实验结果表明,得到的集装箱模型分类器效果理想。
2.提出基于Kalman滤波的彩色模板更新的方法。介绍了基于色彩直方图的模板匹配方法和Kalman滤波的基本原理,给出了色彩直方图的具体算法。对模板内每一个像素值的三个颜色分量使用Kalman滤波进行更新,得到最优的模板图像,然后用模板匹配算法来检测集装箱。此方法与原先的固定模板匹配算法相比,提高了模板匹配的准确度和稳定度。
Currently, using computer vision technology is the direction of container recognition and position in automatic container loading and unloading operations.Two of literatures, surveys and research papers concerning up-to-date techniques of container detection are read and analyzed. Putting up the Container Handling Model System(CHMS), compiling the code in Windows, and carrying out simulations about the two methods on the CHMS. The experiments indicate that the methods of container detection proposed in showing a certain degree of theoretical and practical value. Paper mainly includes the following several aspects:
1.Using container detection method based on AdaBoost learning algorithm, which selects few key haar-like features from a large set of features, to build a robust cascade classifier. The convergence and generalization capability of AdaBoost algorithm as well as the effect of weight-updating approach on the classifier's performance are analysed comprehensively. This paper used the sample storehouse which one founded to carry on the training and get the container classifier. Detecting the container model In a variety of backgrounds and lighting conditions used the classifier. The experimental results show that the effect of container classifier is suitable to detection.
2. Providing an approach to update the template based on a Kalman filter. Intruducing the the basic principles of the template matching based on the color histogram and the Kalman filter, and giving the specific algorithm of the color histogram. The Kalman filter update three component of each pixel value and get the optimum template, and containers are detected by the optimum template. This method improves the accuracy and the stability of template matching.
引文
[1]李智.我国国际集装箱运输的现状及存在问题.国际贸易,1994,(11):39-40.
[2]宋联新.世界集装箱物流大港的启示和枢纽港形成的条件.物流技术,1998,93(6):80-82.
[3]竞明.国际集装箱运输市场回顾与展望.中国远洋航务公告,2006,(2):50-53.
[4]虞卫华.集装箱船舶市场.集装箱化,2004, (2):12-14.
[5]仲伟伫. 2S(GPS+GIS)技术在集装箱港口物流中的应用.第一届国际综合物流学术大会论文集.新加坡南洋理工大学,2001.277-280.
[6]陆青.用于轮胎式龙门集装箱起重机的卫星定位系统移动站.中国专利:01142631.4, 2003-06-25.
[7]蔡芸,孙国正.先进技术在集装箱港口的应用[[J].中国水运:2004 ,(6):45- 46.
[8] Parry H S,Marshall A D,Markham K C. Tracking targets in FLIR images by region template correlation. SPIE[C] 3086 (277):221-232.
[9] Ray.N.Chanda.B. A fast and flexible multiresolutoin snake with a definite termination criterion .Pattern Recognition 2001,(34):1483-1490.
[10] Huang H Y, Qi F H,Zhao X C. Active contour model based on multiscale image analysis[J]. Journal of Computer Research&Development,2000,137 (10) :1240-1245.
[11] Natan Pterfreund. Robust Tracking of Position and Velocity with Kalman Snakes. IEEE Transon Pattern Analysis and Machine Intelligence, 1999, 22(5) :564-569.
[12]唐仁圣.空中目标实时跟踪算法研究及系统设计[硕士学位论文].重庆:重庆大学信号与信息处理,20040425.
[13] Viola P.,Jones M. J.. Robust Real-Time Face Detection. International Journal of Computer Vision 2004,57(2):137-154
[14] Freund Y. Boosting a weak Learning Algorithm by Majority. Information and Computation 1995,121(2): 256-285.
[15] Schapire R E. The Strength of Weak Learnability. Machine Leaming,1990,5(2): 197-227.
[16]宋春雷,王龙,黄琳.学习理论与鲁棒控制.控制理论与应用,2000,17(5):633-636.
[17] Valiant L.G.. A Theory of the Learnable. Communications of the ACM,1984,27(11): 1134-1142.
[18] Kearns M and Valiant L G. Learning Boolean Formulae or Factoring. Technical Report TR-1488,Cambridge, MA: Hard University Aiken Computation Laboratory, 1988
[19] Kearns M. The Computational Complexity of Machine Learning. Cambridge: MIT Press.1990.
[20] Kearns M., Valiant L. G.. Cryptographic Limitations on Learning Boolean Formulae and Finite Automata. Journal of the ACM, 1994,41(1):67-95.
[21] Drucker H., Schapire R. E., Simard P.. Boosting Performance in Neural Networks. International Journal of Pattern Recognition and Artificial Intelligence, 1993, 7(4):705-719
[22] Freund Y., Schapire R. E.. A Decision-Theoretic Generalization of On-Line Learning and an Application to Boosting. Journal of Computer and System Sciences, 1997, 55(1):119-139.
[23] Viola P.,Jones M.. Robust Real-time Object Detection. Second International Workshop on Statistical and Computational Theories of Vision–modeling, Learning, Computing, AND Sampling Vancouver, Canada, July 13, 2001.
[24] Dietterich T.G.. Machine learning research: Four current directions. AI Magazine, 1997, 18(4): 97-136.
[25]王珏,石纯一.机器学习研究.广西师范大学学报(自然科学版),2003,21(2):1-15.
[26] Honik K. Approximation capabilities of multilayer feedforward networks. Neural Networks,1991, (4):551一557.
[27]阎平凡,黄端旭.人工神经-模型.分析与应用.合肥:安徽教育出版社.1993.
[28] Vapnik V. N.著;张学工,译.统计学习理论的实质.北京:清华大学出版社,2000.
[29] Drucker H, Cortes C. Boosting Decision Trees. In: Touretzky D S, Mozer M C, Hasselmo M E,Comes C .eds. Advances in Neural Information Processing .Cambridge: MIT Press 1996,VII: 479一485.
[30] Freund Y, Schapire R E. Experiments with a New Boosting Algorithm. In: Proc of the 13th International Conference on Machine Learning .San Francisco, CA, 1996,48-156.
[31] Schapire R E, Freund Y,Bartlett P,Lee W S. Boosting the Margin: A New Explanation for the Effectiveness of Voting Methods. The Annals of Statistics,1998, 26(5): 1651-1686.
[32] Sehwenk H, Bengio Y. Boosting Neural Netwroks. Neural Computation, 2000 12(8): 1869-1887.
[33] Robert E. Schapire, Yoav Freund, Peter Bartlett and Wee Sun Lee. Boosting the margin: A new explanation for the effectiveness of voting methods. The Annals of Statistics, 1998, 6(5):1651-1686.
[34] Lisa M.Brown. Mutiple View Independent Vehicle Person Classification. IBM T.J Watson Research Center, VSSN04, October 15,2004,New York.
[35]黄涛.模板匹配在图像识别中的应用.云南大学学报,2005,27(5A):327-332.
[36]张恒博,欧宗瑛.一种基于色彩和灰度直方图的图像检索方法.计算机程.2004,30(10): 20-22.
[37]张惠.基于色彩直方图的视频检索问题的研究.[硕士学位论文].北京:北京交通大学通信与信息系统专业.2006.
[38]马艳.基于颜色于模板匹配的人脸检测方法.[硕士学位论文].大连:大连理工大学模式识别与智能系统专业,2006.
[39]杨珺,王继成,刘然.融合空间及邻域信息的色彩直方图检索方法.计算机工程与应用, 2007,43(27):158-160.
[40]汪颖进,张桂林.新的基于卡尔曼滤波的跟踪方法.红外与激光工程.2004,33(5): 505-508.
[41]周宏仁,王敬东,王德培.机动目标跟踪.北京:国防工业出版社,1991.
[42]成光,刘卫东,魏尚俊,张蕊.基于卡尔曼滤波的目标估计和预测方法研究.计算机仿真, 2006,23(1):8-10.
[43]秦勤.雷达目标跟踪的卡尔曼滤波方法的研究.[硕士学位论文].大连:大连海事大学通信与信息系统专业,2006.
[2]宋联新.世界集装箱物流大港的启示和枢纽港形成的条件.物流技术,1998,93(6):80-82.
[3]竞明.国际集装箱运输市场回顾与展望.中国远洋航务公告,2006,(2):50-53.
[4]虞卫华.集装箱船舶市场.集装箱化,2004, (2):12-14.
[5]仲伟伫. 2S(GPS+GIS)技术在集装箱港口物流中的应用.第一届国际综合物流学术大会论文集.新加坡南洋理工大学,2001.277-280.
[6]陆青.用于轮胎式龙门集装箱起重机的卫星定位系统移动站.中国专利:01142631.4, 2003-06-25.
[7]蔡芸,孙国正.先进技术在集装箱港口的应用[[J].中国水运:2004 ,(6):45- 46.
[8] Parry H S,Marshall A D,Markham K C. Tracking targets in FLIR images by region template correlation. SPIE[C] 3086 (277):221-232.
[9] Ray.N.Chanda.B. A fast and flexible multiresolutoin snake with a definite termination criterion .Pattern Recognition 2001,(34):1483-1490.
[10] Huang H Y, Qi F H,Zhao X C. Active contour model based on multiscale image analysis[J]. Journal of Computer Research&Development,2000,137 (10) :1240-1245.
[11] Natan Pterfreund. Robust Tracking of Position and Velocity with Kalman Snakes. IEEE Transon Pattern Analysis and Machine Intelligence, 1999, 22(5) :564-569.
[12]唐仁圣.空中目标实时跟踪算法研究及系统设计[硕士学位论文].重庆:重庆大学信号与信息处理,20040425.
[13] Viola P.,Jones M. J.. Robust Real-Time Face Detection. International Journal of Computer Vision 2004,57(2):137-154
[14] Freund Y. Boosting a weak Learning Algorithm by Majority. Information and Computation 1995,121(2): 256-285.
[15] Schapire R E. The Strength of Weak Learnability. Machine Leaming,1990,5(2): 197-227.
[16]宋春雷,王龙,黄琳.学习理论与鲁棒控制.控制理论与应用,2000,17(5):633-636.
[17] Valiant L.G.. A Theory of the Learnable. Communications of the ACM,1984,27(11): 1134-1142.
[18] Kearns M and Valiant L G. Learning Boolean Formulae or Factoring. Technical Report TR-1488,Cambridge, MA: Hard University Aiken Computation Laboratory, 1988
[19] Kearns M. The Computational Complexity of Machine Learning. Cambridge: MIT Press.1990.
[20] Kearns M., Valiant L. G.. Cryptographic Limitations on Learning Boolean Formulae and Finite Automata. Journal of the ACM, 1994,41(1):67-95.
[21] Drucker H., Schapire R. E., Simard P.. Boosting Performance in Neural Networks. International Journal of Pattern Recognition and Artificial Intelligence, 1993, 7(4):705-719
[22] Freund Y., Schapire R. E.. A Decision-Theoretic Generalization of On-Line Learning and an Application to Boosting. Journal of Computer and System Sciences, 1997, 55(1):119-139.
[23] Viola P.,Jones M.. Robust Real-time Object Detection. Second International Workshop on Statistical and Computational Theories of Vision–modeling, Learning, Computing, AND Sampling Vancouver, Canada, July 13, 2001.
[24] Dietterich T.G.. Machine learning research: Four current directions. AI Magazine, 1997, 18(4): 97-136.
[25]王珏,石纯一.机器学习研究.广西师范大学学报(自然科学版),2003,21(2):1-15.
[26] Honik K. Approximation capabilities of multilayer feedforward networks. Neural Networks,1991, (4):551一557.
[27]阎平凡,黄端旭.人工神经-模型.分析与应用.合肥:安徽教育出版社.1993.
[28] Vapnik V. N.著;张学工,译.统计学习理论的实质.北京:清华大学出版社,2000.
[29] Drucker H, Cortes C. Boosting Decision Trees. In: Touretzky D S, Mozer M C, Hasselmo M E,Comes C .eds. Advances in Neural Information Processing .Cambridge: MIT Press 1996,VII: 479一485.
[30] Freund Y, Schapire R E. Experiments with a New Boosting Algorithm. In: Proc of the 13th International Conference on Machine Learning .San Francisco, CA, 1996,48-156.
[31] Schapire R E, Freund Y,Bartlett P,Lee W S. Boosting the Margin: A New Explanation for the Effectiveness of Voting Methods. The Annals of Statistics,1998, 26(5): 1651-1686.
[32] Sehwenk H, Bengio Y. Boosting Neural Netwroks. Neural Computation, 2000 12(8): 1869-1887.
[33] Robert E. Schapire, Yoav Freund, Peter Bartlett and Wee Sun Lee. Boosting the margin: A new explanation for the effectiveness of voting methods. The Annals of Statistics, 1998, 6(5):1651-1686.
[34] Lisa M.Brown. Mutiple View Independent Vehicle Person Classification. IBM T.J Watson Research Center, VSSN04, October 15,2004,New York.
[35]黄涛.模板匹配在图像识别中的应用.云南大学学报,2005,27(5A):327-332.
[36]张恒博,欧宗瑛.一种基于色彩和灰度直方图的图像检索方法.计算机程.2004,30(10): 20-22.
[37]张惠.基于色彩直方图的视频检索问题的研究.[硕士学位论文].北京:北京交通大学通信与信息系统专业.2006.
[38]马艳.基于颜色于模板匹配的人脸检测方法.[硕士学位论文].大连:大连理工大学模式识别与智能系统专业,2006.
[39]杨珺,王继成,刘然.融合空间及邻域信息的色彩直方图检索方法.计算机工程与应用, 2007,43(27):158-160.
[40]汪颖进,张桂林.新的基于卡尔曼滤波的跟踪方法.红外与激光工程.2004,33(5): 505-508.
[41]周宏仁,王敬东,王德培.机动目标跟踪.北京:国防工业出版社,1991.
[42]成光,刘卫东,魏尚俊,张蕊.基于卡尔曼滤波的目标估计和预测方法研究.计算机仿真, 2006,23(1):8-10.
[43]秦勤.雷达目标跟踪的卡尔曼滤波方法的研究.[硕士学位论文].大连:大连海事大学通信与信息系统专业,2006.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |