基于RF-PCA-改进SVM模型的齿轮故障诊断方法

为提升齿轮故障诊断的正确率,提出了基于随机森林(Random Forest,RF)和主成分分析法(Principal Components Analysis,PCA)对齿轮振动信号进行特征降维处理,并采用粒子群算法(Particle Swarm Optimization,PSO)求解支持向量机(Support Vector Machine,SVM)特征参数最佳取值的齿轮故障诊断模型(RF-PCA-改进SVM模型)。对齿轮箱实例中正常、断齿、齿根开裂、剥落、削尖等9种不同状态进行了验证,证明了RF-PCA-改进SVM模型对齿轮故障诊断的有效性。试验结果表明:通过对比不同诊断模型识别率,证明RF-PCA-改进SVM模型具有更优的齿轮故障识别率,平均达到了99.66%,且计算效率较高;样本数量改变虽然会影响模型正确识别率,但不同的改变方式对识别率影响的程度不同,对于RF-PCA-改进SVM模型,当齿轮状态数据大于40个时即可达到88%以上的正确识别率。     

基于无人机高光谱影像和机器学习的红树林树种精细分类

利用海南省文昌市清澜港红树林保护区的无人机高光谱影像,采用递归特征消除的随机森林算法（Recursive Feature Elimination-Random Forest,RFE-RF）优选植被光谱特征和纹理特征,通过机器学习中的随机森林（Random Forest,RF）和支持向量机（Support Vector Machine,SVM）算法对研究区内的红树林树种进行精细分类,并对比分析和评价分类模型参数设置对总体精度的影响。结果表明：RF分类方法的总体精度为92.70%、 Kappa系数为0.91,与传统的SVM 分类方法相比,RF算法均提高了5类树种的生产者精度和使用者精度,能够有效地对红树林树种进行精细分类,可为种植资源规划和生态环境保护等方面提供技术支持。     

基于图像检测的发动机缸体探伤系统设计及应用

为了实现某型号航空发动机缸体的自动化检测,设计了一套全自动荧光磁粉探伤系统。介绍了系统的结构和实现方式、重点介绍了图像采集和图像处理部分,系统采用Laplacian算子对原始图像进行锐化处理,然后采用Prewitt算子进行边缘提取,最后采用面积和周长两个物理量作为判断产品合格与否的特征值;实际测量显示,系统的测量精度可达0.01 mm,满足发动机缸体检测的要求。     

基于SVM的教学质量监控和评价方法的研究

教学质量是高校的生命线,对教学质量监控及评价数据的加工、分析和利用尤为重要.针对教学质量监控和评价中的问题,文章分析了适合本校的教学质量评价指标,利用SVM理论,结合C#语言和数据库技术,提出了一个基于SVM的教学质量评价方案.对评价数据进行预处理,利用LIBSVM封装技术,对数据进行训练与预测,结果证明,SVM模型能够达到较好的预测精度,评价结果符合要求,具有一定的实用价值.     

深孔镗加工工况监测系统的研究

对深孔镗加工中可利用的工况信号作了分析研究，选出两种信号对其建立ＡＲ模型，并就如何提取特征参数和如何综合集成信号提出了方案并给出了监控系统的结构框图     

基于图像特征的籽棉品级分级模型的研究

为了提高在籽棉收购环节中品级分级的客观性和准确性,籽棉品级分级模型以籽棉图像的特征参数为依据,以BP神经网络为分类算法。依据GB1103-2007中籽棉品级分级标准,提取籽棉图像的白度、黄度、杂质作为特征参数,使用部分样本对BP神经网络进行训练,训练后的BP神经网络对未参加训练的样本进行分级,精度达到81%。实验证明,该模型能够提高籽棉分级的客观性和准确性。     

基于多层集成学习的岩性识别方法

岩性识别是油藏地质解释中的关键问题和难点问题,人工智能特别是机器学习技术的发展和应用为岩性识别问题解决提供了新的技术途径。本文利用支持向量机（Support vector machine,SVM）、多粒度级联森林（Multi-grained cascade forest,GCForest）、随机森林（Random forest,RF）以及XGBoost（eXtreme gradient boosting）等机器学习模型建立一个异构多层集成学习模型,该集成学习模型克服了单一模型对数据集要求高、泛化能力差以及识别精度低等缺点。本文分别利用集成模型和单一模型进行了岩性识别实验。实验结果表明,本文集成模型在岩性分类测试集上平均精度达到96.66%,高于SVM的平均精度75.53%、GCForest的平均精度96.21%、随机森林的平均精度95.06%和XGBoost的平均精度95.77%。该集成模型能有效地用于油藏地质分析中的岩性识别和分类任务,适应性强,识别精度高。     

发动机缸体线性与非线性对比分析

采用线性分析发动机缸体时,在螺栓接触和绑定的位置会出现明显的应力集中点,且计算得到疲劳安全因数远远低于合理值.在考虑材料非线性的基础上对发动机缸体进行非线性分析,发现在螺栓接触和绑定位置处的应力明显降低,计算得到的疲劳安全因数也趋于合理.通过汽油机缸体线性与非线性对比分析可以发现,缸体非线性模型的计算结果更合理.若要改善螺栓孔周边的安全因数,一定要保证螺栓绑定和接触位置接近实际情况,特别在2个面过渡的位置不能绑定,这样可以避免造成应力集中.     

基于核学习算法的驾驶精神疲劳分级研究

为了有效地评测人的驾驶精神疲劳状态,本文提出了一种基于核学习算法的精神疲劳分级方法.该方法首先用多变量自回归模型(MVAR)提取位于前额、顶叶、枕叶共6个通道的多维脑电信号特征组成特征向量.然后用核主分量分析(KPCA)和优化支持向量机(SVM)对基于脑电信号(EEG)的驾驶精神疲劳进行分级.经过对3住受试者在3个状态下的驾驶精神疲劳进行分类,平均分类精度达到89.47%.分析显示,应用KPCA并结合优化SVM方法有效地降低了特征空间的维数,可实现较高精度的驾驶精神疲劳分级.     

基于遗传算法与Shannon熵的故障监控参数优选

研究液体火箭发动机故障诊断中监控参数的优选问题.基于Shannon熵理论提出 了特征参数组所含故障分类信息的理论值及其工程计算方法,证明了故障分类信息与参数相 关性之间的单调降关系,并以此作为特征参数的优选准则,利用改进的遗传算法对某液体火 箭发动机的常见故障进行了特征参数优选,数值实验结果表明所选特征参数合理,且故障分 类器的计算复杂度大大降低而对噪声的鲁棒性大大提高.     

A sensor fusion and support vector machine based approach for recognition of complex machining conditions

During the machining process of thin-walled parts, machine tool wear and work-piece deformation always co-exist, which make the recognition of machining conditions very difficult. Existing machining condition monitoring approaches usually consider only one single condition, i.e., either tool wear or work-piece deformation. In order to close this gap, a machining condition recognition approach based on multi-sensor fusion and support vector machine (SVM) is proposed. A dynamometer sensor and an acceleration sensor are used to collect cutting force signals and vibration signals respectively. Wavelet decomposition is utilized as a signal processing method for the extraction of signal characteristics including means and variances of a certain degree of the decomposed signals. SVM is used as a condition recognition method by using the means and variances of signals as well as cutting parameters as the input vector. Information fusion theory at the feature level is adopted to assist the machining condition recognition. Experiments are designed to demonstrate and validate the feasibility of the proposed approach. A condition recognition accuracy of about 90 % has been achieved during the experiments.     

基于GA的遥感图像目标SVM自动识别

为了高效合理地确定支持矢量机(SVM)的参数,使其对复杂的二值遥感图像目标进行自动识别,采用实值编码遗传算法来实现SVM模型参数的自动选择.与穷举搜索的留一法及随机试凑法相比,采用遗传算法的SVM模型参数选择更简单、更易于实现,并使SVM具有更好的推广能力.二值遥感图像目标的分类识别结果表明,该方法不但可以提高分类识别率,而且显著地缩短了SVM的训练时间.     

Face membership authentication using SVM classification tree generated by membership-based LLE data partition

This paper presents a new membership authentication method by face classification using a support vector machine (SVM) classification tree, in which the size of membership group and the members in the membership group can be changed dynamically. Unlike our previous SVM ensemble-based method, which performed only one face classification in the whole feature space, the proposed method employed a divide and conquer strategy that first performs a recursive data partition by membership-based locally linear embedding (LLE) data clustering, then does the SVM classification in each partitioned feature subset. Our experimental results show that the proposed SVM tree not only keeps the good properties that the SVM ensemble method has, such as a good authentication accuracy and the robustness to the change of members, but also has a considerable improvement on the stability under the change of membership group size.     

Multi-objective uniform design as a SVM model selection tool for face recognition

The primary difficulty of support vector machine (SVM) model selection is heavy computational cost, thus it is difficult for current model selection methods to be applied in face recognition. Model selection via uniform design can effectively alleviate the computational cost, but its drawback is that it adopts a single objective criterion which can not always guarantee the generalization capacity. The sensitivity and specificity as multi-objective criteria have been proved of better performance and can provide a means for obtaining more realistic models. This paper first proposes a multi-objective uniform design (MOUD) search method as a SVM model selection tool, and then applies this optimized SVM classifier to face recognition. Because of replacing single objective criterion with multi-objective criteria and adopting uniform design to seek experimental points that uniformly scatter on whole experimental domain, MOUD can reduce the computational cost and improve the classification ability simultaneously. The experiments are executed on UCI benchmark, and on Yale and CAS-PEAL-R1 face databases. The experimental results show that the proposed method outperforms other model search methods significantly, especially for face recognition.     

Human Face Detection in Digital Video Using SVMEnsemble

This Letter proposes automatic human face detection in digital video using a support vector machine (SVM) ensemble to improve the detection performance. The SVM ensemble consists of several independently trained SVMs using randomly chosen training samples via a bootstrap technique. Next, they are aggregated in order to make a collective decision via a majority voting scheme. Experimental results show that the proposed face detection method using SVM ensemble outperforms conventional methods such as using only single SVM and Multi-Layer Perceptron in terms of classification accuracy, false alarms, and missing rates.     

基于增量学习SVM的人脸识别

为了提高人脸识别率,本文提出了一种增量学习支持矢量机(SVM)人脸识别方法,有效地对SVM的参数进行更新。提出的方法采用高斯概率模型描述SVM的参数统计特征,在无需额外存储训练数据的前提下,采用增量学习SVM的方式实现参数的更新;并通过最小化分类误差准则最大化SVM两类输出值概率分布间的距离。详细的实验以及与现有方法的比较结果表明,提出的识别方法具有更好的识别性能。     

Calibration-based tool condition monitoring for repetitive machining operations

Reliable tool condition monitoring (TCM) system is essential for any machining process in mass production to control the part quality as well as reduce the machine tool downtime and maintenance costs. However, while various research studies have proposed their TCM systems, the complexity in setups with advanced decision-making algorithms and specificity in application to limited cutting conditions continue to complicate the implementation of these systems into practical scenarios. This study develops a very simple and flexible TCM system for repetitive machining operations. The proposed monitoring approach reduces the complexity of monitoring model by considering the important characteristic of repeatability in process which has been commonly found in the mass production scenario and implements the calibration procedure to improve the flexibility of the model application to actual machining processes with complex toolpath designs and variable cutting conditions. The selected cutting tools with specific tool conditions are used in the calibration phase to generate reference signals. In actual repetitive production, the collected signal generated by the cutting tool in each operation is compared with reference signals to identify the most similar condition of the reference tool through the proposed similarity analysis. To validate the performance, the current study demonstrates the application of proposed monitoring approach to monitor the tool wear in repetitive milling operations with complex toolpath, and the predicted tool wear progression is found to be in good agreement with experimental measurements during the machining of multiple parts over the entire tool life.     

基于支持向量机的枪弹外观缺陷识别与分类

为解决枪弹外观缺陷自动分类问题,提出了一种基于支持向量机的枪弹外观缺陷自动识别与分类模型。首先针对枪弹表面缺陷的图像特点,从几何、灰度、纹理三方面进行了特征提取,在此基础上建立了基于支持向量机的枪弹外观缺陷分类模型,并对特征参数进行了优选;研究了支持向量机中惩罚系数和核函数参数对分类器性能的影响;通过实验与基于BP神经网络的枪弹外观缺陷分类器进行了比较,结果表明,在小样本下,基于支持向量机的枪弹外观缺陷分类器性能更好。     

基于离散微粒群优化算法的SVM参数选择

支持向量机(SVM)的学习性能和泛化能力主要取决于参数选择,然而传统的优化算法难以解决此问题。文中通过支持向量的个数建立优化目标函数,采用微粒群优化(PSO)算法对其优化,寻找最优参数。PSO是一种新兴的基于群体智慧的进化算法。实验表明,微粒群优化算法是支持向量机参数选择的有效方法。     

面向不平衡数据集的煤矿监测系统异常数据识别方法

异常数据识别对于煤矿安全监测系统具有重要作用,但安全监测系统中异常数据一般只占数据总量的1%左右,不平衡性是此类数据的固有特点。目前多数机器学习算法在不平衡数据集上的分类预测准确率和灵敏度都相对较差。为了能准确识别异常数据,以煤矿分布式光纤竖井变形监测系统采集的数据为研究对象,提出了一种面向不平衡数据集、基于去重复下采样(RDU)、合成少数类过采样技术(SMOTE)和随机森林(RF)分类算法的煤矿监测系统异常数据识别方法。该方法利用RDU算法对多数类数据进行下采样,去除重复样本;利用SMOTE算法对少数类异常数据进行过采样,通过合成新的异常数据来改善数据集的不平衡性;并利用优化后的数据集训练RF分类算法,得到异常数据识别模型。在6个真实数据集上的对比实验结果表明,该方法的异常数据识别准确率平均值达到99.3%,具有较好的泛化性和较强的鲁棒性。