Wheat ear counting is a prerequisite for the evaluation of wheat yield.A wheat ear counting method based on frequency domain decomposition is proposed in this study to improve the accuracy of wheat yield estimation.Th...Wheat ear counting is a prerequisite for the evaluation of wheat yield.A wheat ear counting method based on frequency domain decomposition is proposed in this study to improve the accuracy of wheat yield estimation.The frequency domain decomposition of wheat ear image is completed by multiscale support value filter(MSVF)combined with improved sampled contourlet transform(ISCT).Support Vector Machine(SVM)is the classic classification and regression algorithm of machine learning.MSVF based on this has strong frequency domain filtering and generalization ability,which can effectively remove the complex background,while the multi-direction characteristics of ISCT enable it to represent the contour and texture information of wheat ears.In order to improve the level of wheat yield prediction,MSVF-ISCT method is used to decompose the ear image in multiscale and multi direction in frequency domain,reduce the interference of irrelevant information,and generate the sub-band image with more abundant information components of ear feature information.Then,the ear feature is extracted by morphological operation and maximum entropy threshold segmentation,and the skeleton thinning and corner detection algorithms are used to count the results.The number of wheat ears in the image can be accurately counted.Experiments show that compared with the traditional algorithms based on spatial domain,this method significantly improves the accuracy of wheat ear counting,which can provide guidance and application for the field of agricultural precision yield estimation.展开更多
This paper presents a voice conversion technique based on bilinear models and introduces the concept of contextual modeling. The bilinear approach reformulates the spectral envelope representation from line spectral f...This paper presents a voice conversion technique based on bilinear models and introduces the concept of contextual modeling. The bilinear approach reformulates the spectral envelope representation from line spectral frequencies feature to a two-factor parameterization corresponding to speaker identity and phonetic information, the so-called style and content factors. This decomposition offers a flexible representation suitable for voice conversion and facilitates the use of efficient training algorithms based on singular value decomposition. In a contextual approach (bilinear) models are trained on subsets of the training data selected on the fly at conversion time depending on the characteristics of the feature vector to be converted. The performance of bilinear models and context modeling is evaluated in objective and perceptual tests by comparison with the popular GMM-based voice conversion method for several sizes and different types of training data.展开更多
A modified matrix enhancement and matrix pencil (MMEMP) method is presented for the scattering centers measurements in step-frequency radar. The method estimates the signal parameter pairs directly unlike the matrix e...A modified matrix enhancement and matrix pencil (MMEMP) method is presented for the scattering centers measurements in step-frequency radar. The method estimates the signal parameter pairs directly unlike the matrix enhancement and matrix pencil (MEMP) method which contains an additional step to pair the parameters related to each dimension. The downrange and crossrange expressions of the scattering centers are deduced, as well as the range ambiguities, from the point of view of MMEMP method. Compared with the Fourier transform method, the numerical simulation shows that both the resolution and precision of the MMEMP method are higher than those of the Fourier method. The processing results of the real measured data for three cylinders prove the above conclusions further.展开更多
射频干扰(Radio Frequency Interference,RFI)会对高频地波雷达有用回波产生较大影响。本文提出了一种慢时域射频干扰抑制方法,首先利用频谱监测数据实现射频干扰的分段检测,而后基于射频干扰在慢时域的短时相干性、强距离相关性和方向...射频干扰(Radio Frequency Interference,RFI)会对高频地波雷达有用回波产生较大影响。本文提出了一种慢时域射频干扰抑制方法,首先利用频谱监测数据实现射频干扰的分段检测,而后基于射频干扰在慢时域的短时相干性、强距离相关性和方向特性,在常规高阶奇异值分解(Higher-Order Singular Value Decomposition,HOSVD)方法的基础上,结合训练张量三种展开模式矩阵的特点,利用左、右奇异矩阵包含的频率信息实现对干扰子空间的准确估计,进而实现对射频干扰的分段消除。仿真和实测数据的处理结果都表明,该方法可以有效检测并消除射频干扰,提高了数据批处理的运算效率。展开更多
近年来,机器学习在计算机视觉中取得了许多突破性的研究进展.然而,已训练好的学习模型难以直接应用于相似但具有不同数据分布特征的其它学习任务中.域自适应技术通过抽取源域与目标域数据之间的公共特征,来实现把源域中学习到的知识迁...近年来,机器学习在计算机视觉中取得了许多突破性的研究进展.然而,已训练好的学习模型难以直接应用于相似但具有不同数据分布特征的其它学习任务中.域自适应技术通过抽取源域与目标域数据之间的公共特征,来实现把源域中学习到的知识迁移至目标域,从而避免针对目标域的训练数据收集和模型训练代价.但是,现有的视觉域自适应方法大都无法处理高阶的特征数据,一般都是通过简单的向量化操作将高阶张量特征转换成高维一阶向量特征.这不仅会破坏高阶特征数据内部的结构信息,而且还会增加算法的计算复杂度.为了解决上述问题,本文在保持原有张量特征结构不变的条件下,利用张量乘操作,将视觉域自适应问题抽象为求解源域和目标域的共同张量子空间以及源域和目标域特征在该共同张量子空间上投影的多变量优化问题.然后,利用张量奇异值分解和交替方向乘子法,提出一种基于张量奇异值分解的视觉域自适应方法(Visual domain Adaptation method based on TEnsor Singular value decomposition,VATES),以实现上述多变量优化问题的迭代求解.文中证明了正交张量子空间约束条件下源域与目标域表征误差最小化问题的可解性问题,并求得了相应的解析解.在公开数据集Office-Caltech-10、Office31、ImageNet-VOC2007上与17个基线模型进行对比实验.结果表明本文所提出的方法与经典的机器学习方法、非深度域自适应方法、深度域自适应方法以及张量域自适应方法相比,在无标签目标域上的图像分类精度分别提高了10.6%~43.9%、0.7%~31.1%、0.7%~24.8%以及5.7%~34.9.同时,算法的运行效率也提高了40.5%~74.3%,显著优于所对比的基线方法.实验分析也表明,VATES方法的目标域分类精度会随着所选用神经网络特征抽取能力的增强而逐渐提升.展开更多
在直流电路系统中,电弧故障是引起电气火灾的主要原因,有效的线路电弧故障检测能够确保线路的安全运行和设备的可靠工作。为解决上述问题,该文引入奇异值分解法(Singular value decomposition,SVD)对采集到的样本数据进行特征向量提取...在直流电路系统中,电弧故障是引起电气火灾的主要原因,有效的线路电弧故障检测能够确保线路的安全运行和设备的可靠工作。为解决上述问题,该文引入奇异值分解法(Singular value decomposition,SVD)对采集到的样本数据进行特征向量提取。首先,设计直流串联电弧故障实验平台,对电弧故障特性进行分析;其次,介绍SVD的特征向量提取原理和支持向量机识别机制;最后,对实验结果进行分析,进一步验证所提检测方法的可行性和适用性。展开更多
With the development of wireless communication technology, the electromagnetic environment has become more and more complex. Conventional signal identification methods are difficult to accurately identify illegal devi...With the development of wireless communication technology, the electromagnetic environment has become more and more complex. Conventional signal identification methods are difficult to accurately identify illegal devices. However, electromagnetic signals have an unavoidable device-specific characteristic unintentionally generated by a transmitter, appearing in the form of an Un Intentional Modulation(UIM), namely Radio Frequency Fingerprint(RFF). RFFs can be used to uniquely identify an emitter to match a received signal with its source. In this paper, the authors propose a novel RFF scheme to separate UIM part from the original signals from the time and frequency domain, and then utilize non-Gaussian measuring tools to extract a set of dimensionreduced secondary features. Additionally, Singular Value Reconstruction(SVR) is developed to extract UIM in the frequency spectrum. In time domain, a curve-fitting residual method is proposed to extract the UIM on the estimated instantaneous phase based on Maximum Likelihood Estimator(MLE). Various aspects of the proposed method are evaluated, including identification accuracy under various Signal-to-Noise Ratio(SNR) conditions, energy relationships between the UIM and the whole signal, and sensitivity to training set size. Compared with other methods, experimental results based on real-world signals prove that the proposed method has remarkable performance and high practicability.展开更多
基金National Natural Science Foundation of China(61672032)National Key Research and Development Program of China(2016YFD0800904)+1 种基金Anhui Provincial Science and Technology Project(16030701091)The Open Research Fund of National Engineering Research Center for Agro-Ecological Big Data Analysis&Application,Anhui University(AE2018009).
文摘Wheat ear counting is a prerequisite for the evaluation of wheat yield.A wheat ear counting method based on frequency domain decomposition is proposed in this study to improve the accuracy of wheat yield estimation.The frequency domain decomposition of wheat ear image is completed by multiscale support value filter(MSVF)combined with improved sampled contourlet transform(ISCT).Support Vector Machine(SVM)is the classic classification and regression algorithm of machine learning.MSVF based on this has strong frequency domain filtering and generalization ability,which can effectively remove the complex background,while the multi-direction characteristics of ISCT enable it to represent the contour and texture information of wheat ears.In order to improve the level of wheat yield prediction,MSVF-ISCT method is used to decompose the ear image in multiscale and multi direction in frequency domain,reduce the interference of irrelevant information,and generate the sub-band image with more abundant information components of ear feature information.Then,the ear feature is extracted by morphological operation and maximum entropy threshold segmentation,and the skeleton thinning and corner detection algorithms are used to count the results.The number of wheat ears in the image can be accurately counted.Experiments show that compared with the traditional algorithms based on spatial domain,this method significantly improves the accuracy of wheat ear counting,which can provide guidance and application for the field of agricultural precision yield estimation.
文摘This paper presents a voice conversion technique based on bilinear models and introduces the concept of contextual modeling. The bilinear approach reformulates the spectral envelope representation from line spectral frequencies feature to a two-factor parameterization corresponding to speaker identity and phonetic information, the so-called style and content factors. This decomposition offers a flexible representation suitable for voice conversion and facilitates the use of efficient training algorithms based on singular value decomposition. In a contextual approach (bilinear) models are trained on subsets of the training data selected on the fly at conversion time depending on the characteristics of the feature vector to be converted. The performance of bilinear models and context modeling is evaluated in objective and perceptual tests by comparison with the popular GMM-based voice conversion method for several sizes and different types of training data.
文摘A modified matrix enhancement and matrix pencil (MMEMP) method is presented for the scattering centers measurements in step-frequency radar. The method estimates the signal parameter pairs directly unlike the matrix enhancement and matrix pencil (MEMP) method which contains an additional step to pair the parameters related to each dimension. The downrange and crossrange expressions of the scattering centers are deduced, as well as the range ambiguities, from the point of view of MMEMP method. Compared with the Fourier transform method, the numerical simulation shows that both the resolution and precision of the MMEMP method are higher than those of the Fourier method. The processing results of the real measured data for three cylinders prove the above conclusions further.
文摘射频干扰(Radio Frequency Interference,RFI)会对高频地波雷达有用回波产生较大影响。本文提出了一种慢时域射频干扰抑制方法,首先利用频谱监测数据实现射频干扰的分段检测,而后基于射频干扰在慢时域的短时相干性、强距离相关性和方向特性,在常规高阶奇异值分解(Higher-Order Singular Value Decomposition,HOSVD)方法的基础上,结合训练张量三种展开模式矩阵的特点,利用左、右奇异矩阵包含的频率信息实现对干扰子空间的准确估计,进而实现对射频干扰的分段消除。仿真和实测数据的处理结果都表明,该方法可以有效检测并消除射频干扰,提高了数据批处理的运算效率。
文摘近年来,机器学习在计算机视觉中取得了许多突破性的研究进展.然而,已训练好的学习模型难以直接应用于相似但具有不同数据分布特征的其它学习任务中.域自适应技术通过抽取源域与目标域数据之间的公共特征,来实现把源域中学习到的知识迁移至目标域,从而避免针对目标域的训练数据收集和模型训练代价.但是,现有的视觉域自适应方法大都无法处理高阶的特征数据,一般都是通过简单的向量化操作将高阶张量特征转换成高维一阶向量特征.这不仅会破坏高阶特征数据内部的结构信息,而且还会增加算法的计算复杂度.为了解决上述问题,本文在保持原有张量特征结构不变的条件下,利用张量乘操作,将视觉域自适应问题抽象为求解源域和目标域的共同张量子空间以及源域和目标域特征在该共同张量子空间上投影的多变量优化问题.然后,利用张量奇异值分解和交替方向乘子法,提出一种基于张量奇异值分解的视觉域自适应方法(Visual domain Adaptation method based on TEnsor Singular value decomposition,VATES),以实现上述多变量优化问题的迭代求解.文中证明了正交张量子空间约束条件下源域与目标域表征误差最小化问题的可解性问题,并求得了相应的解析解.在公开数据集Office-Caltech-10、Office31、ImageNet-VOC2007上与17个基线模型进行对比实验.结果表明本文所提出的方法与经典的机器学习方法、非深度域自适应方法、深度域自适应方法以及张量域自适应方法相比,在无标签目标域上的图像分类精度分别提高了10.6%~43.9%、0.7%~31.1%、0.7%~24.8%以及5.7%~34.9.同时,算法的运行效率也提高了40.5%~74.3%,显著优于所对比的基线方法.实验分析也表明,VATES方法的目标域分类精度会随着所选用神经网络特征抽取能力的增强而逐渐提升.
文摘在直流电路系统中,电弧故障是引起电气火灾的主要原因,有效的线路电弧故障检测能够确保线路的安全运行和设备的可靠工作。为解决上述问题,该文引入奇异值分解法(Singular value decomposition,SVD)对采集到的样本数据进行特征向量提取。首先,设计直流串联电弧故障实验平台,对电弧故障特性进行分析;其次,介绍SVD的特征向量提取原理和支持向量机识别机制;最后,对实验结果进行分析,进一步验证所提检测方法的可行性和适用性。
基金supported by the Program for Innovative Research Groups of the Hunan Provincial Natural Science Foundation of China(No.2019JJ10004)。
文摘With the development of wireless communication technology, the electromagnetic environment has become more and more complex. Conventional signal identification methods are difficult to accurately identify illegal devices. However, electromagnetic signals have an unavoidable device-specific characteristic unintentionally generated by a transmitter, appearing in the form of an Un Intentional Modulation(UIM), namely Radio Frequency Fingerprint(RFF). RFFs can be used to uniquely identify an emitter to match a received signal with its source. In this paper, the authors propose a novel RFF scheme to separate UIM part from the original signals from the time and frequency domain, and then utilize non-Gaussian measuring tools to extract a set of dimensionreduced secondary features. Additionally, Singular Value Reconstruction(SVR) is developed to extract UIM in the frequency spectrum. In time domain, a curve-fitting residual method is proposed to extract the UIM on the estimated instantaneous phase based on Maximum Likelihood Estimator(MLE). Various aspects of the proposed method are evaluated, including identification accuracy under various Signal-to-Noise Ratio(SNR) conditions, energy relationships between the UIM and the whole signal, and sensitivity to training set size. Compared with other methods, experimental results based on real-world signals prove that the proposed method has remarkable performance and high practicability.
