2006年 第32卷 第3期
2006, 32(3): 321-328.
摘要:
Aiming at the coupling characteristic between the two groups of electromagnets embedded in the module of the maglev train, a nonlinear decoupling controller is designed. The module is modeled as a double-electromagnet system, and based on some reasonable assumptions its nonlinear mathematical model, a MIMO coupling system, is derived. To realize the linearization and decoupling from the input to the output, the model is linearized exactly by means of feedback linearization, andan equivalent linear decoupling model is obtained. Based on the linear model, a nonlinear suspension controller is designed using state feedback. Simulations and experiments show that the controller can effectually solve the coupling problem in double-electromagnet suspension system.
Aiming at the coupling characteristic between the two groups of electromagnets embedded in the module of the maglev train, a nonlinear decoupling controller is designed. The module is modeled as a double-electromagnet system, and based on some reasonable assumptions its nonlinear mathematical model, a MIMO coupling system, is derived. To realize the linearization and decoupling from the input to the output, the model is linearized exactly by means of feedback linearization, andan equivalent linear decoupling model is obtained. Based on the linear model, a nonlinear suspension controller is designed using state feedback. Simulations and experiments show that the controller can effectually solve the coupling problem in double-electromagnet suspension system.
2006, 32(3): 322-328.
摘要:
Aiming at the coupling characteristic between the two groups of electromagnets embedded in the module of the maglev train, a nonlinear decoupling controller is designed. The module is modeled as a double-electromagnet system, and based on some reasonable assumptions its nonlinear mathematical model, a MIMO coupling system, is derived. To realize the linearization and decoupling from the input to the output, the model is linearized exactly by means of feedback linearization, and an equivalent linear decoupling model is obtained. Based on the linear model, a nonlinear suspension controller is designed using state feedback. Simulations and experiments show that the controller can effectually solve the coupling problem in double-electromagnet suspension system.
Aiming at the coupling characteristic between the two groups of electromagnets embedded in the module of the maglev train, a nonlinear decoupling controller is designed. The module is modeled as a double-electromagnet system, and based on some reasonable assumptions its nonlinear mathematical model, a MIMO coupling system, is derived. To realize the linearization and decoupling from the input to the output, the model is linearized exactly by means of feedback linearization, and an equivalent linear decoupling model is obtained. Based on the linear model, a nonlinear suspension controller is designed using state feedback. Simulations and experiments show that the controller can effectually solve the coupling problem in double-electromagnet suspension system.
2006, 32(3): 329-336.
摘要:
The adaptive fault-tolerant control scheme of dynamic nonlinear system based on the credit assigned fuzzy CMAC neural network is presented. The proposed learning approach uses the learned times of addressed hypercubes as the credibility, the amounts of correcting errors are proportional to the inversion of the learned times of addressed hypercubes. With this idea, the learning speed can indeed be improved. Based on the improved CMAC learning approach and using the sliding control technique, the effective control law reconfiguration strategy is presented. Thesystem stability and performance are analyzed under failure scenarios. The numerical simulation demonstrates the effectiveness of the improved CMAC algorithm and the proposed fault-tolerant controller.
The adaptive fault-tolerant control scheme of dynamic nonlinear system based on the credit assigned fuzzy CMAC neural network is presented. The proposed learning approach uses the learned times of addressed hypercubes as the credibility, the amounts of correcting errors are proportional to the inversion of the learned times of addressed hypercubes. With this idea, the learning speed can indeed be improved. Based on the improved CMAC learning approach and using the sliding control technique, the effective control law reconfiguration strategy is presented. Thesystem stability and performance are analyzed under failure scenarios. The numerical simulation demonstrates the effectiveness of the improved CMAC algorithm and the proposed fault-tolerant controller.
2006, 32(3): 337-344.
摘要:
对基于绝对二次曲线和基于绝对二次曲面的两类摄像机自标定方法的鲁棒性进行了分析,并从矩阵条件数出发,通过大量仿真实验对两类方法进行了定量比较．实验结果表明, 基于绝对二次曲线的摄像机自标定方法的系数矩阵的条件数一般小于基于绝对二次曲面方法的系数矩阵的条件数．另外,当常数因子有误差时,基于绝对二次曲面方法的系数矩阵条件数的变化一般比基于绝对二次曲线方法系数矩阵条件数的变化更剧烈．上述二点表明,基于绝对二次曲线的自标定方法的鲁棒性一般要优于基于绝对二次曲面的自标定方法．上述结论与文献中的一些实验观察正好相佐．
对基于绝对二次曲线和基于绝对二次曲面的两类摄像机自标定方法的鲁棒性进行了分析,并从矩阵条件数出发,通过大量仿真实验对两类方法进行了定量比较．实验结果表明, 基于绝对二次曲线的摄像机自标定方法的系数矩阵的条件数一般小于基于绝对二次曲面方法的系数矩阵的条件数．另外,当常数因子有误差时,基于绝对二次曲面方法的系数矩阵条件数的变化一般比基于绝对二次曲线方法系数矩阵条件数的变化更剧烈．上述二点表明,基于绝对二次曲线的自标定方法的鲁棒性一般要优于基于绝对二次曲面的自标定方法．上述结论与文献中的一些实验观察正好相佐．
2006, 32(3): 345-352.
摘要:
无模型控制律在炼油、化工、电力、轻工等领域应用获得了良好的效果．应用的方式许多是串级形式．将对无模型控制律的串级形式进行分析,指出这种形式是对具有干扰的大时滞系统控制的一种有效的方法,给出了在加热炉、氢氮比控制中成功应用的实例．
无模型控制律在炼油、化工、电力、轻工等领域应用获得了良好的效果．应用的方式许多是串级形式．将对无模型控制律的串级形式进行分析,指出这种形式是对具有干扰的大时滞系统控制的一种有效的方法,给出了在加热炉、氢氮比控制中成功应用的实例．
2006, 32(3): 353-359.
摘要:
提出了一种基于自适应特征与多级反馈模型的新颖的字符分割方法,对文字图像质量与中英文混排格式有较好的自适应能力．该方法的主要思想就是将一个分割过程分成很多层,每层都会由一个主要特征来指导字符分割与中英文预分类,然后将分割层的结果反馈至当前分割层或前面的分割层,并指导下一层的分割．该方法将字符分割、中英文预分类和字符识别这三者进行了很好的融合,大大提高了字符分割与识别的正确率．
提出了一种基于自适应特征与多级反馈模型的新颖的字符分割方法,对文字图像质量与中英文混排格式有较好的自适应能力．该方法的主要思想就是将一个分割过程分成很多层,每层都会由一个主要特征来指导字符分割与中英文预分类,然后将分割层的结果反馈至当前分割层或前面的分割层,并指导下一层的分割．该方法将字符分割、中英文预分类和字符识别这三者进行了很好的融合,大大提高了字符分割与识别的正确率．
2006, 32(3): 360-367.
摘要:
提出了一种新的生物特征识别模式--手指背关节皮纹识别．利用自主设计的采集装置获得手背图像,由Canny算子和滑动窗分割并定位手指背关节皮纹．在识别时,首先对要检验的两背关节皮纹进行快速配准,然后用两种方法识别,并对两种方法进行了比较,一是基于相关分类器的识别,一是基于复Gabor小波变换的识别．后者是利用复Gabor小波提取背关节皮纹特征,并利用二进制编码得到特征码,以两背关节皮纹特征码的Hammming距离为判据,检验两者是否为同一模式．试验结果表明:手指背关节皮纹具有较高的唯一性,可以用作身份认证,在等错误率情况下,基于相关分类器的识别准确率高达98.04%,基于Gabor小波变换的识别准确率为94.61%．而后者比前者的识别速度要快得多．
提出了一种新的生物特征识别模式--手指背关节皮纹识别．利用自主设计的采集装置获得手背图像,由Canny算子和滑动窗分割并定位手指背关节皮纹．在识别时,首先对要检验的两背关节皮纹进行快速配准,然后用两种方法识别,并对两种方法进行了比较,一是基于相关分类器的识别,一是基于复Gabor小波变换的识别．后者是利用复Gabor小波提取背关节皮纹特征,并利用二进制编码得到特征码,以两背关节皮纹特征码的Hammming距离为判据,检验两者是否为同一模式．试验结果表明:手指背关节皮纹具有较高的唯一性,可以用作身份认证,在等错误率情况下,基于相关分类器的识别准确率高达98.04%,基于Gabor小波变换的识别准确率为94.61%．而后者比前者的识别速度要快得多．
2006, 32(3): 368-377.
摘要:
粒子群优化算法在优化问题中体现出良好的性能,但目前还没有对其运动特性,尤其是参数的选择与当粒子群体陷入局部极值点导致的早熟收敛情况的详细分析．分析了PSO算法中的三种粒子模型(Gbest,Pbest,Commom模型)的运动特性,给出了Gbest模型和Pbest 模型在没有新息获取时,单信息条件下的最大搜索空间．进一步证明了在减少了Lipschitz条件约束的条件下,Common模型渐进稳定的充分条件,将算法中惯量因子的取值范围扩大到 (-1,1),并从物理上进行了解释．
粒子群优化算法在优化问题中体现出良好的性能,但目前还没有对其运动特性,尤其是参数的选择与当粒子群体陷入局部极值点导致的早熟收敛情况的详细分析．分析了PSO算法中的三种粒子模型(Gbest,Pbest,Commom模型)的运动特性,给出了Gbest模型和Pbest 模型在没有新息获取时,单信息条件下的最大搜索空间．进一步证明了在减少了Lipschitz条件约束的条件下,Common模型渐进稳定的充分条件,将算法中惯量因子的取值范围扩大到 (-1,1),并从物理上进行了解释．
2006, 32(3): 378-385.
摘要:
将用于两类分类的最大散度差鉴别准则推广为多类最大散度差鉴别准则,并建立了基于该准则的一种新的人脸表示方法.基于多类最大散度差鉴别准则的人脸表示方法有效避免了传统鉴别分析方法在人脸特征提取时通常面临的小样本模式识别问题.在国际标准人脸图像数据库ORL、Yale以及FERET上的实验结果表明,与Fisherfaces、Eigenfaces、正交补空间、零空间等人脸特征提取方法相比,新的人脸表示方法具有一定的优势.
将用于两类分类的最大散度差鉴别准则推广为多类最大散度差鉴别准则,并建立了基于该准则的一种新的人脸表示方法.基于多类最大散度差鉴别准则的人脸表示方法有效避免了传统鉴别分析方法在人脸特征提取时通常面临的小样本模式识别问题.在国际标准人脸图像数据库ORL、Yale以及FERET上的实验结果表明,与Fisherfaces、Eigenfaces、正交补空间、零空间等人脸特征提取方法相比,新的人脸表示方法具有一定的优势.
2006, 32(3): 386-392.
摘要:
针对基于主成分分析识别人脸存在计算复杂、不能准确地估计训练图像的协方差矩阵等问题,提出了一种基于描述特征的人脸识别算法(Expressive feature face recognitionalgorithm, EFFRA).该算法用训练图像的右奇异向量代替PCA求解的子空间的基向量,避免了将人脸图像转换成图像向量,明显降低了计算复杂性.进一步研究发现,EFFRA提取的每一个主成分向量中含有冗余,在此基础上,利用PCA实现了EFFRA的简化算法(MEFFRA),在ORL和Essex数据库上的实验结果表明,EFFRA及MEFFRA明显优于特征脸算法,MEFFRA的识别精度略好于EFFRA,但明显减少了对存储空间的需求.
针对基于主成分分析识别人脸存在计算复杂、不能准确地估计训练图像的协方差矩阵等问题,提出了一种基于描述特征的人脸识别算法(Expressive feature face recognitionalgorithm, EFFRA).该算法用训练图像的右奇异向量代替PCA求解的子空间的基向量,避免了将人脸图像转换成图像向量,明显降低了计算复杂性.进一步研究发现,EFFRA提取的每一个主成分向量中含有冗余,在此基础上,利用PCA实现了EFFRA的简化算法(MEFFRA),在ORL和Essex数据库上的实验结果表明,EFFRA及MEFFRA明显优于特征脸算法,MEFFRA的识别精度略好于EFFRA,但明显减少了对存储空间的需求.
2006, 32(3): 393-399.
摘要:
研究了遗传程序设计(GP)算法中适应度函数的光滑拟合问题,结合LAMs(Linear association memorys)方法和HJ(Hook和Jeevs)方法两种方法,估计GP树数值权值,以减少GP树适应度值评价的计算代价.光滑拟合的好坏关键取决于调整参数的选择.提出了一种选择调整参数的新方法,同时,给出了两个数学例子,并与广义交叉实验B-样条函数仿真比较验证.
研究了遗传程序设计(GP)算法中适应度函数的光滑拟合问题,结合LAMs(Linear association memorys)方法和HJ(Hook和Jeevs)方法两种方法,估计GP树数值权值,以减少GP树适应度值评价的计算代价.光滑拟合的好坏关键取决于调整参数的选择.提出了一种选择调整参数的新方法,同时,给出了两个数学例子,并与广义交叉实验B-样条函数仿真比较验证.
2006, 32(3): 400-410.
摘要:
现代过程工业正逐渐倚重于生产小批量、多品种、高附加值产品的间歇过程.基于多元统计模型的过程监测是保障生产安全和产品质量的重要工具.从间歇过程独特的数据特性出发,将现有的多元统计建模方法进行合理的分类,简要回顾了各类方法的起源、发展及延伸的历程.除了阐述每种方法的基本原理,还详细讨论了各种方法的适用背景,相互关联及优缺点等内容,并对这一领域中依然存在的问题以及研究前景给出中肯的评述.
现代过程工业正逐渐倚重于生产小批量、多品种、高附加值产品的间歇过程.基于多元统计模型的过程监测是保障生产安全和产品质量的重要工具.从间歇过程独特的数据特性出发,将现有的多元统计建模方法进行合理的分类,简要回顾了各类方法的起源、发展及延伸的历程.除了阐述每种方法的基本原理,还详细讨论了各种方法的适用背景,相互关联及优缺点等内容,并对这一领域中依然存在的问题以及研究前景给出中肯的评述.
2006, 32(3): 411-416.
摘要:
A new hierarchical switching control system of multiple models based on robust control theory is designed for some plant with large uncertainties. The model set and controller set are designed by robust control theory and the characteristics of robust control system are taken into account. A new kind of switching index function by estimating uncertainty is designed. Furthermore,stability of the closed system is analyzed by the small gain theorem in the sense of exponentially weighted L2 norm. And simulation is done on a plant with both parameter uncertainty and unmodeled dynamics. Both theoretical analysis and simulation results show that this new hierarchical switching control system can control the plant with large uncertainties effectively and has good performance of tracking and stability.
A new hierarchical switching control system of multiple models based on robust control theory is designed for some plant with large uncertainties. The model set and controller set are designed by robust control theory and the characteristics of robust control system are taken into account. A new kind of switching index function by estimating uncertainty is designed. Furthermore,stability of the closed system is analyzed by the small gain theorem in the sense of exponentially weighted L2 norm. And simulation is done on a plant with both parameter uncertainty and unmodeled dynamics. Both theoretical analysis and simulation results show that this new hierarchical switching control system can control the plant with large uncertainties effectively and has good performance of tracking and stability.
2006, 32(3): 417-421.
摘要:
A multiscale principal component analysis method is proposed for sensor fault detection and identification. After decomposition of sensor signal by wavelet transform, the coarse-scale coef-ficients from the sensors with strong correlation are employed to establish the principal component analysis model. A moving window is designed to monitor data from each sensor using the model.For the purpose of sensor fault detection and identification, the data in the window is decomposed with wavelet transform to acquire the coarse-scale coefficients firstly, and the square prediction error is used to detect the failure. Then the sensor validity index is introduced to identify faulty sensor,which provides a quantitative identifying index rather than qualitative contrast given by the approach with contribution. Finally, the applicability and effectiveness of the proposed method is illustrated by sensors of industrial boiler.
A multiscale principal component analysis method is proposed for sensor fault detection and identification. After decomposition of sensor signal by wavelet transform, the coarse-scale coef-ficients from the sensors with strong correlation are employed to establish the principal component analysis model. A moving window is designed to monitor data from each sensor using the model.For the purpose of sensor fault detection and identification, the data in the window is decomposed with wavelet transform to acquire the coarse-scale coefficients firstly, and the square prediction error is used to detect the failure. Then the sensor validity index is introduced to identify faulty sensor,which provides a quantitative identifying index rather than qualitative contrast given by the approach with contribution. Finally, the applicability and effectiveness of the proposed method is illustrated by sensors of industrial boiler.
2006, 32(3): 422-427.
摘要:
Abstract The control of underactuated mechanical systems is very complex for the loss of its control inputs. The model of underactuated mechanical systems in a potential field is built with Lagrangian method and its structural properties are analyzed in detail. A stable control approach is proposed for the class of underactuated mechanical systems. This approach is applied to an unde ractuated double-pendulum-type overhead crane and the simulation results illustrate the correctness of dynamics analysis and validity of the proposed control algorithm.
Abstract The control of underactuated mechanical systems is very complex for the loss of its control inputs. The model of underactuated mechanical systems in a potential field is built with Lagrangian method and its structural properties are analyzed in detail. A stable control approach is proposed for the class of underactuated mechanical systems. This approach is applied to an unde ractuated double-pendulum-type overhead crane and the simulation results illustrate the correctness of dynamics analysis and validity of the proposed control algorithm.
2006, 32(3): 428-432.
摘要:
Symmetric workpiece localization algorithms combine alternating optimization and linearization. The iterative variables are partitioned into two groups. Then simple optimization approaches can be employed for each subset of variables, where optimization of configuration variables is simplified as a linear least-squares problem (LSP). Convergence of current symmetric localization algorithms is discussed firstly. It is shown that simply taking the solution of the LSP as start of the next iteration may result in divergence or incorrect convergence. Therefore in our enhanced algorithms, line search is performed along the solution of the LSP in order to find a better point reducing the value of objective function. We choose this point as start of the next iteration. Better convergence is verified by numerical simulation. Besides, imposing boundary constraints on the LSP proves to be another efficient way.
Symmetric workpiece localization algorithms combine alternating optimization and linearization. The iterative variables are partitioned into two groups. Then simple optimization approaches can be employed for each subset of variables, where optimization of configuration variables is simplified as a linear least-squares problem (LSP). Convergence of current symmetric localization algorithms is discussed firstly. It is shown that simply taking the solution of the LSP as start of the next iteration may result in divergence or incorrect convergence. Therefore in our enhanced algorithms, line search is performed along the solution of the LSP in order to find a better point reducing the value of objective function. We choose this point as start of the next iteration. Better convergence is verified by numerical simulation. Besides, imposing boundary constraints on the LSP proves to be another efficient way.
2006, 32(3): 433-437.
摘要:
This paper focuses on the fast rate fault detection filter (FDF) problem for a class of multirate sampled-data (MSD) systems. A lifting technique is used to convert such an MSD system into a linear time-invariant discrete-time one and an unknown input observer (UIO) is considered as FDF to generate residual. The design of FDF is formulated as an H∞ optimization problem and a solvable condition as well as an optimal solution are derived. The causality of the residual generator can be guaranteed so that the fast rate residual can be implemented via inverse lifting. A numerical example is included to demonstrate the feasibility of the obtained results.
This paper focuses on the fast rate fault detection filter (FDF) problem for a class of multirate sampled-data (MSD) systems. A lifting technique is used to convert such an MSD system into a linear time-invariant discrete-time one and an unknown input observer (UIO) is considered as FDF to generate residual. The design of FDF is formulated as an H∞ optimization problem and a solvable condition as well as an optimal solution are derived. The causality of the residual generator can be guaranteed so that the fast rate residual can be implemented via inverse lifting. A numerical example is included to demonstrate the feasibility of the obtained results.
2006, 32(3): 438-443.
摘要:
This paper describes the synthesis of robust and non-fragile H∞ state feedback controllers for a class of uncertain jump linear systems with Markovian jumping parameters and state multiplicative noises. Under the assumption of a complete access to the norm-bounds of the system uncertainties and controller gain variations, sufficient conditions on the existence of robust stochastic stability and γ-disturbanee attenuation H∞ property are presented. A key feature of this scheme is that the gain matrices of controller are only based on lt, the observed projection of the current regime rt.
This paper describes the synthesis of robust and non-fragile H∞ state feedback controllers for a class of uncertain jump linear systems with Markovian jumping parameters and state multiplicative noises. Under the assumption of a complete access to the norm-bounds of the system uncertainties and controller gain variations, sufficient conditions on the existence of robust stochastic stability and γ-disturbanee attenuation H∞ property are presented. A key feature of this scheme is that the gain matrices of controller are only based on lt, the observed projection of the current regime rt.
2006, 32(3): 444-449.
摘要:
为了改善电动车的可操作性和稳定性,本文进行电动车制动防抱死控制系统的研究.首先针对汽车纵向四轮模型,设计自适应鲁棒控制器.所提出的自适应控制器和自适应鲁棒控制律不仅保证了闭环系统的稳定性,而且实现了所期望的性能.然后通过试验结果证明了控制算法的有效性.尽管道路条件的变化不同,电动车制动防抱死控制系统表现出满意的控制性能.
为了改善电动车的可操作性和稳定性,本文进行电动车制动防抱死控制系统的研究.首先针对汽车纵向四轮模型,设计自适应鲁棒控制器.所提出的自适应控制器和自适应鲁棒控制律不仅保证了闭环系统的稳定性,而且实现了所期望的性能.然后通过试验结果证明了控制算法的有效性.尽管道路条件的变化不同,电动车制动防抱死控制系统表现出满意的控制性能.
2006, 32(3): 450-455.
摘要:
对带回滞驱动的一类单输入单输出的非线性不确定系统,本文采用Prandtl-Ishlinskii模型描述回滞特性,采用反步递推设计方法,实现自适应控制器的设计.仿真结果说明控制方法的有效性.
对带回滞驱动的一类单输入单输出的非线性不确定系统,本文采用Prandtl-Ishlinskii模型描述回滞特性,采用反步递推设计方法,实现自适应控制器的设计.仿真结果说明控制方法的有效性.
2006, 32(3): 456-461.
摘要:
网格技术是一种新型的分布计算技术,致力于解决复杂度很高的新应用问题.随着全球半导体生产规模的日益扩大,半导体生产线的优化调度问题成为学术界及工程界研究的热点.半导体生产线具有许多特殊的特点,诸如生产规模大、工件数量多、随机性大、加工成本高、高度的可重入性等,这些特点决定了原有的调度策略已不能满足半导体生产线的要求.鉴于网格技术在处理设备可扩展性和资源平衡性上的优势,主要研究将网格技术的思想用于半导体生产线的调度中.利用网格计算中的负载向量和失衡因子的概念,来控制半导体生产线上各加工机器处工件块的规模以及投料规模.通过优化算法的调度,使得半导体生产线的各加工设备负载得到平衡,设备的生产效率提高,缩短加工周期,从而达到优化生产线的目的.
网格技术是一种新型的分布计算技术,致力于解决复杂度很高的新应用问题.随着全球半导体生产规模的日益扩大,半导体生产线的优化调度问题成为学术界及工程界研究的热点.半导体生产线具有许多特殊的特点,诸如生产规模大、工件数量多、随机性大、加工成本高、高度的可重入性等,这些特点决定了原有的调度策略已不能满足半导体生产线的要求.鉴于网格技术在处理设备可扩展性和资源平衡性上的优势,主要研究将网格技术的思想用于半导体生产线的调度中.利用网格计算中的负载向量和失衡因子的概念,来控制半导体生产线上各加工机器处工件块的规模以及投料规模.通过优化算法的调度,使得半导体生产线的各加工设备负载得到平衡,设备的生产效率提高,缩短加工周期,从而达到优化生产线的目的.
2006, 32(3): 462-469.
摘要:
在ESPRT检测方法的基础上,结合新息过程给出了一种新的变点检测方法.利用ESPRT方法检测经新息模型产生的新息过程,实现了模型参数变点的非参数检测,扩展了ESPRT方法在实际应用中的适用范围.将该方法应用于长输管道泄漏故障监测时,利用基于BP神经网络的非线性时间序列预测方法建立了管道泄漏监测系统的新息模型以及泄漏检测模型.将上述监测方法应用于实验泄漏水管道,能在线实时有效地发现泄漏.
在ESPRT检测方法的基础上,结合新息过程给出了一种新的变点检测方法.利用ESPRT方法检测经新息模型产生的新息过程,实现了模型参数变点的非参数检测,扩展了ESPRT方法在实际应用中的适用范围.将该方法应用于长输管道泄漏故障监测时,利用基于BP神经网络的非线性时间序列预测方法建立了管道泄漏监测系统的新息模型以及泄漏检测模型.将上述监测方法应用于实验泄漏水管道,能在线实时有效地发现泄漏.
2006, 32(3): 470-474.
摘要:
研究了傅立叶变换、不变矩的原理及特点,提出基于幅值谱与不变矩的特征提取方法,并应用于中厚板的表面缺陷自动分类.从现场在线采集中厚板的表面图像,将每幅表面图像划分成128×128大小的子图像,对子图像进行傅立叶变换得到子图像的幅值谱,再对幅值谱图像求Hu不变矩,将不变矩作为特征量,通过这种方法提取的特征向量不仅具有平移、旋转不变性,并且具有抗噪、抑制光照不均的优点.将本文方法得到的特征量作为基于LVQ神经网络的分类器输入,对缺陷样本进行学习和分类,结果表明,这些特征量适用于中厚板表面缺陷的分类,识别率达81.5％.
研究了傅立叶变换、不变矩的原理及特点,提出基于幅值谱与不变矩的特征提取方法,并应用于中厚板的表面缺陷自动分类.从现场在线采集中厚板的表面图像,将每幅表面图像划分成128×128大小的子图像,对子图像进行傅立叶变换得到子图像的幅值谱,再对幅值谱图像求Hu不变矩,将不变矩作为特征量,通过这种方法提取的特征向量不仅具有平移、旋转不变性,并且具有抗噪、抑制光照不均的优点.将本文方法得到的特征量作为基于LVQ神经网络的分类器输入,对缺陷样本进行学习和分类,结果表明,这些特征量适用于中厚板表面缺陷的分类,识别率达81.5％.
2006, 32(3): 475-480.
摘要:
提出了一种基于内禀模态(Intrinsic mode functions,简称IMFs)奇异值分解和支持向量机(Support vector machine,简称SVM)的故障诊断方法.采用经验模态分解(Empirical mode decomposition,简称EMD)方法对旋转机械故障振动信号进行分解,将得到的若干个内禀模态分量自动形成初始特征向量矩阵,然后对该矩阵进行奇异值分解,提取其奇异值作为故障特征向量,并进一步根据支持向量机分类器的输出结果来判断旋转机械的工作状态和故障类型.对齿轮振动信号的分析结果表明,即使在小样本情况下,基于内禀模态奇异值分解和支持向量机的故障诊断方法仍能有效地识别齿轮的工作状态和故障类型.
提出了一种基于内禀模态(Intrinsic mode functions,简称IMFs)奇异值分解和支持向量机(Support vector machine,简称SVM)的故障诊断方法.采用经验模态分解(Empirical mode decomposition,简称EMD)方法对旋转机械故障振动信号进行分解,将得到的若干个内禀模态分量自动形成初始特征向量矩阵,然后对该矩阵进行奇异值分解,提取其奇异值作为故障特征向量,并进一步根据支持向量机分类器的输出结果来判断旋转机械的工作状态和故障类型.对齿轮振动信号的分析结果表明,即使在小样本情况下,基于内禀模态奇异值分解和支持向量机的故障诊断方法仍能有效地识别齿轮的工作状态和故障类型.
