Application of wavelet transform in feature extraction and pattern recognition of wideband echoes

A novel approach to extract edge features from wideband echo is proposed. The set of extracted features not only represents the echo waveform in a concise way, but also is sufficient and well suited for classification of non-stationary echo data from objects with different property.The feature extraction is derived from the Discrete Dyadic Wavlet Transform (DDWT) of the echo through the undecimated algorithm. The motivation we use the DDWT is that it is time-shift-invariant which is beneficial for localization of edge, and the wavelet coefficients at larger scale represent the main shape feature of echo, i.e. edge, and the noise and modulated high-frequency components are reduced with scale increased. Some experimental results using real data which contain 144 samples from 4 classes of lake bottoms with different sediments are provided. The results show that our approach is a prospective way to represent wideband echo for reliable recognition of nonstationary echo with great variability.

Face Representation Using Combined Method of Gabor Filters, Wavelet Transformation and DCV and Recognition Using RBF

An efficient face representation is a vital step for a successful face recognition system. Gabor features are known to be effective for face recognition. The Gabor features extracted by Gabor filters have large dimensionality. The feature of wavelet transformation is feature reduction. Hence, the large dimensional Gabor features are reduced by wavelet transformation. The discriminative common vectors are obtained using the within-class scatter matrix method to get a feature representation of face images with enhanced discrimination and are classified using radial basis function network. The proposed system is validated using three face databases such as ORL, The Japanese Female Facial Expression (JAFFE) and Essex Face database. Experimental results show that the proposed method reduces the number of features, minimizes the computational complexity and yielded the better recognition rates.

Wavelet packet based feature extraction and recognition of license plate characters

To study the characteristics of license plate characters recognition,this paper proposes a method for fea- ture extraction of license plate characters based on two-dimensional wavelet packet.We decompose license plate character images with two dimensional-wavelet packet and search for the optimal wavelet packet basis.This paper pre- sents a criterion of searching for the optimal wavelet packet basis,and a practical algorithm.The obtained optimal wavelet packet basis is used as the feature of license plate character,and a BP neural network is used to classify the character.The test- ing results show that the proposed method achieved higher recognition rate than the traditional methods.

Handwritten Character Recognition Using Multiresolution Technique and Euclidean Distance Metric

In the present paper, the problem of handwritten character recognition has been tackled with multiresolution technique using discrete wavelet transform (DWT) and Euclidean distance metric (EDM). The technique has been tested and found to be more accurate and faster. Characters is classified into 26 pattern classes based on appropriate properties. Features of the handwritten character images are extracted by DWT used with appropriate level of multiresolution technique, and then each pattern class is characterized by a mean vector. Distances from input pattern vector to all the mean vectors are computed by EDM. Minimum distance determines the class membership of input pattern vector. The proposed method provides good recognition accuracy of 90% for handwritten characters even with fewer samples.

FACE RECOGNITION BASED ON WAVELET-CURVELET-FRACTAL TECHNIQUE

In this paper,a novel face recognition method,named as wavelet-curvelet-fractal technique,is proposed. Based on the similarities embedded in the images,we propose to utilize the wave-let-curvelet-fractal technique to extract facial features. Thus we have the wavelet’s details in diagonal,vertical,and horizontal directions,and the eight curvelet details at different angles. Then we adopt the Euclidean minimum distance classifier to recognize different faces. Extensive comparison tests on dif-ferent data sets are carried out,and higher recognition rate is obtained by the proposed technique.

基于面积加权GWT-GFT的水声目标识别

由于海洋环境的复杂性,水声目标的识别具有很大的挑战性。为解决这类复杂环境下特征提取的问题,提出了一种基于面积加权的图小波变换-图傅里叶变换(GWT-GFT)的分析方法。在完成数据预处理后,为了能够凸显顶点之间的关系,提出了一种新的基于顶点三角形面积的加权方法来构建图信号;构建好的图信号通过GWT分解为多尺度图分量;然后,利用GFT将这些分量从图域变换到特征值谱域进行分析;在此基础上,提取各分量特征值谱的特征;最后,利用基于高斯核函数的支持向量机(SVM)对获取的特征向量进行分类。基于水声信号ShipsEar数据库,采用5折交叉验证方法进行验证。与现有的其它方法相比,所提的模型以36个特征在376656个样本上取得了97.22%的准确率,证明了该分析方法的有效性和鲁棒性。

基于深度融合模型的气膜密封端面状态识别方法

气膜密封装置是工业领域应用广泛的一种密封技术,其可靠的密封性能对于设备正常运行至关重要。气膜密封装置的动静密封环接触端面相对运动会产生摩擦,摩擦过程会产生复杂的声发射信号,这些信号往往隐含密封端面运行状况的重要信息。采用传统的方法往往难以准确识别和分类这些微弱的特征信号,因此需要开发更高精度的故障诊断方法。针对机械密封动、静环端面摩擦状态难以识别这一问题,以气膜密封装置为研究对象,提出了一种基于深度融合模型的气膜密封端面状态识别方法。首先,采用声发射传感器及采集设备,对密封端面的声发射信号进行了采集;其次,利用小波包变换方法对采集到的信号进行了滤波处理,并提取了时域和频域的微弱特征;然后,将深度随机森林(DRF)作为分类层融入卷积神经网络(CNN)形成了融合模型,对预先处理过的密封装置运行状态的特征信息进行了识别和分类;最后,根据实验的泄漏量,使用混淆矩阵和受试者工作曲线分析了两种模型的特征提取能力。研究结果表明:CNN-DRF融合模型对于密封端面声发射信号的两种特征识别精度分别为96%和98%,与传统的CNN模型相比,其可以充分提取信号特征信息,具有更出色的故障诊断能力。

基于EWT-SVM的雨量识别方法

为了从雨声信号中识别出雨量的大小,提出了一种基于经验小波变换和支持向量机的雨量识别算法。对于采集到的雨声信号先进行去噪,接着对信号进行经验小波变换分解,分解后得到数个经验小波函数分量,然后通过Matlab编程对各个经验小波函数分量进行特征提取,在时域和频域范围内组成评价特征矩阵,最后通过SVM对特征矩阵进行分类识别。通过仿真实验发现,对于同一个信号,经验小波函数相较于经验模态分解有更好的自适应性并且克服了经验模态分解的混叠现象和端点效应。实验结果表明基于经验小波变换和支持向量机的雨量识别方法在雨量识别领域具有良好的效果,研究方法为雨量识别、智能雨量计的发展奠定了良好的基础。

Optimized Complex Power Quality Classifier Using One vs. Rest Support Vector Machines

Nowadays, power quality issues are becoming a significant research topic because of the increasing inclusion of very sensitive devices and considerable renewable energy sources. In general, most of the previous power quality classification techniques focused on single power quality events and did not include an optimal feature selection process. This paper presents a classification system that employs Wavelet Transform and the RMS profile to extract the main features of the measured waveforms containing either single or complex disturbances. A data mining process is designed to select the optimal set of features that better describes each disturbance present in the waveform. Support Vector Machine binary classifiers organized in a “One Vs Rest” architecture are individually optimized to classify single and complex disturbances. The parameters that rule the performance of each binary classifier are also individually adjusted using a grid search algorithm that helps them achieve optimal performance. This specialized process significantly improves the total classification accuracy. Several single and complex disturbances were simulated in order to train and test the algorithm. The results show that the classifier is capable of identifying >99% of single disturbances and >97% of complex disturbances.

Stand-Alone Intelligent Voice Recognition System

In this paper, an expert system for security based on biometric human features that can be obtained without any contact with the registering sensor is presented. These features are extracted from human’s voice, so the system is called Voice Recognition System (VRS). The proposed system?consists of a combination of three stages: signal pre-processing, features extraction by using?Wavelet Packet Transform (WPT) and features matching by using Artificial Neural Networks (ANNs). The features vectors are formed after two steps: firstly, decomposing the speech signal at level 7 with Daubechies 20-tap (db20), secondly, the energy corresponding to each WPT node is calculated which collected to form a features vector. One hundred twenty eight features vector for each speaker was fed to the Feed Forward Back-propagation Neural Network (FFBPNN). The data used in this paper are drawn from the English Language Speech Database for Speaker Recognition (ELSDSR) database which composes of audio files for training and other files for testing. The performance of the proposed system is evaluated by using the test files. Our results showed that the rate of correct recognition of the proposed system is about 100% for training files and 95.7% for one testing file for each speaker from the ELSDSR database. The proposed method showed efficiency results were better than the well-known Mel Frequency Cepstral Coefficient (MFCC) and the Zak transform.

WSD-SVM在工作面底板破坏深度微震事件自动识别中的应用

为解决煤矿微震事件识别中效率低、精度低、可靠性差的问题,将小波散射分解变换与支持向量机相结合,构建微震事件的WSD-SVM智能识别模型。首先,通过小波散射分解变换将微震监测数据分解成高、低频部分,并计算得到小波散射系数,构成散射特征矩阵;然后,选择70%的数据输入支持向量机模型进行训练,用得到的识别模型对其余30%的数据进行测试验证,获得识别结果。将山西保德煤矿某工作面微震监测时序数据作为实例,结果表明:WSD-SVM模型能够自动识别全部6个微震事件,用时1.651 s;而传统STA/LTA算法虽然仅用时0.731 s,但未能有效识别出其中的3个低信噪比事件,WSD-SVM模型的自动识别精度高于STA/LTA算法模型识别的精度,但需要较长的计算时长。小波散射分解变换方法的引入能够有效实现监测数据降维,大幅提高识别精度,为微震事件的自动识别提供了新思路。

基于EWT-WOA-SVM的三电平逆变器早期故障诊断

针对NPC型三电平逆变器中绝缘栅双极型集体管(IGBT)早期故障信号特征不明显,导致故障特征提取难度大,准确率低等问题,提出了一种基于经验小波变换(EWT)与支持向量机(SVM)相结合的故障诊断方法.该方法对逆变器早期故障进行分析,提取不同故障情况下的三相线电压作为原始特征信号,利用EWT对信号进行故障特征提取;针对EWT方法中需要人为设置分割层数的问题,采用自适应频谱分割方法根据采集到的原始信号自适应确定分割层数;利用鲸鱼优化算法(WOA)优化SVM的诊断方法对故障进行分类.通过仿真实验,并与其他方法进行对比,进一步验证了所提方法的有效性和准确性.

基于APSO-SSVM的异步电动机转子故障诊断

基于信号分析的异步电动机的转子断条与偏心故障诊断方法中,常用传统的电机电流信号特征分析(MCSA)方法。由于采样频率偏低、强大的基波旁瓣效应等因素的影响,会导致特征频率成分被淹没、难以量化故障程度等问题。因此,提出了一种基于自适应粒子群优化逐序支持向量机(APSO-SSVM)的异步电动机故障诊断方法。首先,利用经验小波变换(EWT)对原始信号进行滤波;然后,对滤波后的信号进行特征提取并输入到SSVM诊断模型中;最后,通过APSO算法确定各次序下SVM模型的最佳超参数,从而实现转子断条数量的精确故障诊断。

结合小波与线性映射的人脸图像特征提取算法

提出一种将遗传算法优化和小波与线性映射相结合的特征提取算法。首先对图像数据进行若干层次的小波特征提取,将分解后得到的低频分量采用线性映射算法进行进一步特征提取降维,采用遗传算法对小波分解尺度进行优化,寻找最优的分解尺度以最大化图像识别准确率与最小化特征提取过程的耗时。实验使用支持向量机(SVM)在ORL人脸数据集上比较识别精度和特征提取过程耗时,图像识别准确率比传统方法提高了3%,而特征提取过程耗时则降低了35%~54%,比较试验结果证明图像识别效果切实得到了提高。

基于深度学习的图像文字识别技术在数字化审计中的应用

通过整合的数字化审计平台获取所需审计信息,对所获信息中的图像展开小波变换分解,滤除审计图像噪声的同时提升其质量。结合基于Canny算子的边缘检测与区域分割技术分割完成预处理的审计图像,将完成分割的审计图像输入卷积神经网络,通过卷积层卷积处理审计图像提取审计图像文字特征。经下采样层累积卷积结果得出审计图像文字特征映射图,通过全连接网络判定所提取审计图像文字特征,输出识别结果后通过基于Relief算法的多特征融合方法改进卷积神经网络识别结果。实验结果表明,该方法具有较好的噪声滤除与分割效果,能够精准、清晰识别审计图像中的文字信息。

基于典型相关分析的组合特征抽取及脸像鉴别

利用典型相关分析的思想,提出了一种基于特征级融合的组合特征抽取新方法.首先,抽取同一模式的两组特征矢量,给出描述两组特征矢量之间相关性的判据准则函数;然后依此准则抽取它们的典型相关特征,构成有效鉴别特征矢量用于识别.该方法巧妙地将两组特征矢量之间的相关性特征作为有效判别信息,既达到了信息融合之目的,又消除了特征之间的信息冗余,为两组特征融合用于分类识别提供了新的思路.此外,从理论上进一步剖析了所提出的方法之所以能有效地用于识别的内在本质.在Yale和ORL标准人脸数据库上的实验结果证实了所提算法的有效性和稳定性,而且识别率大大高于用单一特征进行识别的结果.

基于提升小波包变换和集成支持矢量机的早期故障智能诊断

为了解决机电设备早期故障难以正确识别的问题,有效地提高分类的准确率,提出一种基于提升小波包变换和集成支持矢量机的早期故障智能诊断新方法。首先,该方法采用提升策略构造基于冲击故障信号特征的双正交小波,借助提升小波包变换提取信号的敏感频带特征,从而通过对敏感频带中的小波包系数进行包络解调分析检测出故障特征频率。其次,通过距离评估技术从原始信号和小波包系数的统计特征中选取最优特征集。最后,将最优特征输入到集成支持矢量机中,实现对不同故障类型的识别。将该方法应用于滚动轴承的早期故障诊断中,测试结果表明,该方法能够有效地提取故障特征,具有比单一支持矢量机更好的分类性能,故障诊断准确率更高。

基于小波及非线性预测的轴承故障诊断方法

在非线性时间序列预测研究的基础上,提出非线性预测效果的特征提取方法。首先对采集到的足够长轴承数据采用小波变换进行消噪处理及边界延拓,使其满足预测需要的无限长、无噪声的条件,这样延迟时间取任意值均能重构原系统相空间;然后采用基于可预测性的选取嵌入维数的方法确定轴承各种状态信号的嵌入维数,进行相空间重构。应用实验结果表明:该方法提取的特征值能明显地区分轴承各种状态信号,且对数据分段长度的稳定性好,可以作为识别轴承故障的一种新途径。

采用小波分析与神经网络技术的局部放电统计识别方法

小波变换技术适合于局部放电(PD)的检测与识别,但所用小波或提取的特征量不合适时,识别效果不理想。因此,测量了6种典型绝缘缺陷模型产生的144组PD脉冲数据,用基数B样条导数小波成功地从这些实测数据的极强电磁干扰中检测出了PDs,得到了相应的三维统计分布图与各种二维统计分布图,提出了一种新的小波分解方法处理这些PDs对应的统计分布图,并在小波变换域定义了一组与传统方法完全不同的新特征量来识别这些PDs。 以新定义的特征量组成输入矢量,用基于BP算法的前馈型神经网络,对6种典型PDs及加上4种混合PDs组成的共10种PDs进行了识别测试,识别效果远远优于现有方法水平。

基于电力变压器振动信息的绕组形变诊断方法

变压器绕组的机械形变为变压器安全运行埋下事故隐患,为此,提出一种基于振动法的变压器绕组机械状态诊断方法。诊断过程中,通过负载电流与振动信号基频进行拟合匹配初步判定绕组状态;采用小波包变换对变压器绕组不同状态下振动信号进行分析,以振动信号能谱熵作为特征输入向量;利用改进后的多分类支持向量机对特征向量进行训练与测试,实现了变压器绕组不同状态的分类诊断。通过对S11-M-500/35型实际变压器绕组不同状态下进行负载试验,采集对应机械及电气参量数据,用所提出诊断方法对变压器绕组机械状态进行诊断,结果表明:在准确判断绕组正常及故障状态的同时,故障类型诊断结果准确率达到96.78%以上,从而验证所提出诊断方法应用于变压器绕组故障诊断的有效性和准确性。