Cross-Dimension Attentive Feature Fusion Network for Unsupervised Time-Series Anomaly Detection

Time series anomaly detection is crucial in various industrial applications to identify unusual behaviors within the time series data.Due to the challenges associated with annotating anomaly events,time series reconstruction has become a prevalent approach for unsupervised anomaly detection.However,effectively learning representations and achieving accurate detection results remain challenging due to the intricate temporal patterns and dependencies in real-world time series.In this paper,we propose a cross-dimension attentive feature fusion network for time series anomaly detection,referred to as CAFFN.Specifically,a series and feature mixing block is introduced to learn representations in 1D space.Additionally,a fast Fourier transform is employed to convert the time series into 2D space,providing the capability for 2D feature extraction.Finally,a cross-dimension attentive feature fusion mechanism is designed that adaptively integrates features across different dimensions for anomaly detection.Experimental results on real-world time series datasets demonstrate that CAFFN performs better than other competing methods in time series anomaly detection.

Brain Time Stack图像融合技术在CT中的应用

目的:分析Brain Time Stack图像融合技术在CT中的应用。方法:选取2021年3月—2022年9月衡水市第四人民医院收治的50例CT检查患者作为研究对象。所有患者进行CT检查并进行Brain Time Stack后处理。比较四组不同部位CT值、标准差(SD)、信噪比(SNR)。比较四组图像主观质量评分。分析不同部位CT值、SD、SNR与图像主观质量评分的相关性。结果:B组的延髓、额叶灰质、额叶白质、小脑内侧、小脑外侧、颞肌肌肉CT值明显低于A组;C组的延髓、脑室、额叶白质、小脑内侧、小脑外侧、颞肌肌肉CT值高于A组;D组延髓、额叶灰质、颞肌肌肉CT值明显低于A组,脑室、额叶白质、小脑外侧CT值明显高于A组;C组延髓、额叶灰质、额叶白质、小脑内侧、小脑外侧、颞肌肌肉CT值明显高于B组;D组延髓、脑室、额叶白质、小脑内侧、小脑外侧、颞肌肌肉CT值明显高于B组;D组延髓、额叶灰质、额叶白质、小脑内侧、小脑外侧、颞肌肌肉CT值明显低于C组;D组脑室CT值明显高于C组,差异有统计学意义(P<0.05)。B组、C组、D组延髓、脑室、额叶灰质、额叶白质、小脑内侧、小脑外侧、颞肌肌肉SD值明显低于A组;C组延髓、脑室、额叶白质、小脑内侧、小脑外侧、颞肌肌肉SD值均明显高于B组;C组额叶灰质SD明显低于B组;D组延髓、脑室、额叶灰质、额叶白质、小脑内侧、小脑外侧、肌肉SD均明显低于B组、C组,差异有统计学意义(P<0.05)。B组、C组、D组延髓、脑室、额叶灰质、额叶白质、小脑内侧、小脑外侧、颞肌肌肉SNR均明显高于A组;C组、D组延髓、额叶灰质、额叶白质、小脑内侧、小脑外侧、颞肌肌肉SNR值明显高于B组;C组、D组脑室SNR明显低于B组;D组延髓、脑室、额叶灰质、额叶白质、小脑内侧、小脑外侧、颞肌肌肉SNR明显高于C组,差异有统计学意义(P<0.05)。D组图像主观质量评分最高,差异有统计学意义(P<0.05)。延髓、脑室、额叶灰质、额叶白质、小脑内侧、小脑外侧及颞肌肌肉SD与主观质量评分呈明显负相关,SNR与主观质量评分间呈明显正相关,差异有统计学意义(P<0.05)。结论:利用Brain Time Stack图像融合技术对头部CT扫描检查图像处理,动脉期结合前一期及后一期的图像数据在处理后具有更好的质量和更少的噪音。

Prediction of Gas Emission Based on Infor-mation Fusion and Chaotic Time Series

In order to make more exact predictions of gas emissions, information fusion and chaos time series are com- bined to predict the amount of gas emission in pits. First, a multi-sensor information fusion frame is established. The frame includes a data level, a character level and a decision level. Functions at every level are interpreted in detail in this paper. Then, the process of information fusion for gas emission is introduced. On the basis of those data processed at the data and character levels, the chaos time series and neural network are combined to predict the amount of gas emission at the decision level. The weights of the neural network are gained by training not by manual setting, in order to avoid subjectivity introduced by human intervention. Finally, the experimental results were analyzed in Matlab 6.0 and prove that the method is more accurate in the prediction of the amount of gas emission than the traditional method.

Comparison of Spatiotemporal Fusion Models for Producing High Spatiotemporal Resolution Normalized Difference Vegetation Index Time Series Data Sets

It has a great significance to combine multi-source with different spatial resolution and temporal resolution to produce high spatiotemporal resolution Normalized Difference Vegetation Index (NDVI) time series data sets. In this study, four spatiotemporal fusion models were analyzed and compared with each other. The models included the spatial and temporal adaptive reflectance model (STARFM), the enhanced spatial and temporal adaptive reflectance fusion model (ESTARFM), the flexible spatiotemporal data fusion model (FSDAF), and a spatiotemporal vegetation index image fusion model (STVIFM). The objective of is to: 1) compare four fusion models using Landsat-MODIS NDVI image from the Banan district, Chongqing Province;2) analyze the prediction accuracy quantitatively and visually. Results indicate that STVIFM would be more suitable to produce NDVI time series data sets.

INFORMATION FUSION STEADY-STATE WHITE NOISE DECONVOLUTION ESTIMATORS WITH TIME-DELAYED MEASUREMENTS AND COLORED MEASUREMENT NOISES

White noise deconvolution or input white noise estimation problem has important appli-cation backgrounds in oil seismic exploration,communication and signal processing.By the modern time series analysis method,based on the Auto-Regressive Moving Average(ARMA) innovation model,under the linear minimum variance optimal fusion rules,three optimal weighted fusion white noise deconvolution estimators are presented for the multisensor systems with time-delayed measurements and colored measurement noises.They can handle the input white noise fused filtering,prediction and smoothing problems.The accuracy of the fusers is higher than that of each local white noise estimator.In order to compute the optimal weights,the formula of computing the local estimation error cross-covariances is given.A Monte Carlo simulation example for the system with 3 sensors and the Bernoulli-Gaussian input white noise shows their effectiveness and performances.

Deep learning framework for time series classification based on multiple imaging and hybrid quantum neural networks

Time series classification(TSC)has attracted a lot of attention for time series data mining tasks and has been applied in various fields.With the success of deep learning(DL)in computer vision recognition,people are starting to use deep learning to tackle TSC tasks.Quantum neural networks(QNN)have recently demonstrated their superiority over traditional machine learning in methods such as image processing and natural language processing,but research using quantum neural networks to handle TSC tasks has not received enough attention.Therefore,we proposed a learning framework based on multiple imaging and hybrid QNN(MIHQNN)for TSC tasks.We investigate the possibility of converting 1D time series to 2D images and classifying the converted images using hybrid QNN.We explored the differences between MIHQNN based on single time series imaging and MIHQNN based on the fusion of multiple time series imaging.Four quantum circuits were also selected and designed to study the impact of quantum circuits on TSC tasks.We tested our method on several standard datasets and achieved significant results compared to several current TSC methods,demonstrating the effectiveness of MIHQNN.This research highlights the potential of applying quantum computing to TSC and provides the theoretical and experimental background for future research.

Multi-rate sensor fusion-based adaptive discrete finite-time synergetic control for flexible-joint mechanical systems

This paper proposes an adaptive discrete finite-time synergetic control （ADFTSC） scheme based on a multi-rate sensor fusion estimator for flexible-joint mechanical systems in the presence of unmeasured states and dynamic uncertainties. Multi-rate sensors are employed to observe the system states which cannot be directly obtained by encoders due to the existence of joint flexibilities. By using an extended Kalman filter （EKF）, the finite-time synergetic controller is designed based on a sensor fusion estimator which estimates states and parameters of the mechanical system with multi-rate measurements. The proposed controller can guarantee the finite-time convergence of tracking errors by the theoretical derivation. Simulation and experimental studies are included to validate the effectiveness of the proposed approach.

基于数据驱动的分布式光伏发电功率预测方法研究进展

从综述的角度,以分布式光伏系统为对象,分析了功率预测技术的发展情况、存在的难点以及主要影响因素,梳理了应用数据驱动方法实现功率准确预测的技术路线。再以空间相关性、历史出力功率以及气象等影响因素为切入点,梳理了光伏系统数据驱动的功率预测研究现状,分析其相应的数据增强、时空图信息以及特征融合的手段,讨论了技术的优缺点。最后给出了功率预测数据驱动方法研究方向和发展建议。

Image Fusion Real-time System Based on FPGA and Multi-DSP

In order to solve complex algorithm that is difficult to achieve real-time processing of Multiband image fusion within large amount of data, a real-time image fusion system based on FPGA and multi-DSP is designed. Five-band image acquisition, image registration, image fusion and display output can be done within the system which uses FPGA as the main processor and the other three DSP as an algorithm processor. Making full use of Flexible and high-speed characteristics of FPGA, while an image fusion algorithm based on multi-wavelet transform is optimized and applied to the system. The final experimental results show that the frame rate of 15 Hz, with a resolution of 1392 × 1040 of the five-band image can be used by the system to complete processing within 41ms.

一种时频尺度下的多元短期电力负荷组合预测方法

随机因素的增加导致电力负荷数据成分日渐复杂,使短期负荷预测的难度逐渐增大。针对该问题,提出一种时频尺度下的时间卷积网络与多元线性回归相融合的组合预测模型。利用自适应噪声完备集合经验模态分解(complete ensemble empirical mode decomposition with adaptive noise,CEEMDAN)在时频域上将负荷数据分解为若干个频率特征不同的本征模态分量,在模糊熵准则下聚类为随机项和趋势项。采用皮尔逊系数从诸多影响因素中筛选出与电力负荷高度相关的特征,鉴于小时间尺度分析更易于挖掘局部细节特征,分别构建了随机项与趋势项的细颗粒度特征集。利用具有强非线性处理能力的时间卷积网络(temporal convolutional network,TCN)去预测随机项,利用结构简单及线性拟合效果好的多元线性回归(multiplelinearregression,MLR)去预测趋势项,将二者的预测结果进行叠加重构后获得最终预测值。在新加坡和比利时两组数据集上的实验结果证明:所提模型具有较高的预测精度、较好的泛化性能及鲁棒性。

基于多尺度特征信息融合的时间序列异常检测

目前大多数的时间序列都缺少相应的异常标签,且现有基于重构的异常检测算法不能很好地捕获到多维数据间复杂的潜在相关性和时间依赖性,为了构建特征丰富的时间序列,提出一种多尺度特征信息融合的异常检测模型。该模型首先通过卷积神经网络对滑动窗口内的不同序列进行特征卷积来获取不同尺度下的局部上下文信息。然后,利用Transformer中的位置编码对卷积后的时间序列窗口进行位置嵌入,增强滑动窗口中每一个时间序列和邻近序列之间的位置联系,并引入时间注意力获取数据在时间维度上的自相关性,并进一步通过多头自注意力自适应地为窗口内不同时间序列分配不同的权重。最后,对反卷积过程中上采样得到的窗口数据与不同尺度下得到的局部特征和时间上下文信息进行逐步融合,从而准确重构原始时间序列,并将重构误差作为最终的异常得分进行异常判定。实验结果表明,所构建模型在SWaT和SMD数据集上与基线模型相比F1分数均有所提升。在数据维度高且均衡性较差的WADI数据集上与GDN模型相比F1分数降低了1.66%。

小尺度靶丸冲击波调控的冲击波测量技术优化及应用

激光聚变研究中,冲击波调控技术是实现靶丸压缩过程的熵增调谐,保证高性能内爆的关键实验技术.本文在十万焦耳激光装置上首次实现了0.375 mm半径小尺度内爆靶丸下双台阶辐射驱动的高精度冲击波调控实验测量.针对小靶丸下任意反射面速度干涉仪(VISAR)诊断有效反射区域不足的问题,通过建立的球形反射面VISAR图像光强的理论计算方法,提出了利用锁孔(keyhole)锥反射效应提升VISAR诊断空间区域的实验技术路线,使得小靶丸尺度下有效VISAR数据区域提升了近3倍.在实验中首次获得了整形内爆实验条件下低温液氘靶的冲击波测量实验数据,实现了高精度冲击波调控实验测量.实验发现,小时空尺度内爆设计条件下,由于反射冲击波的作用,激光参数的较小偏差都会对冲击波追赶后的传输行为产生显著影响,揭示了我国当前小靶丸尺度下高性能整形内爆物理过程中冲击波传输的多因素敏感性,以及冲击波调控实验对于内爆设计验证的重要性.本文提出的小靶丸冲击波调控实验技术,不仅为我国十万焦耳激光装置上整形脉冲下熵增调谐实验的开展提供了技术基础,也对基于球汇聚压缩的超高压物理研究具有重要意义.

基于多源信息融合的船舶电气设备状态识别方法

为可靠掌握船舶电气设备状态,保证设备的运行安全,提出多源信息融合的船舶电气设备状态识别方法。采用时间序列模型检测并修正船舶电气设备多源历史数据中的连续异常数据和独立异常数据;基于联合卡尔曼滤波算法融合修正后的电气设备多源历史数据,依据融合后的多源数据训练谱聚类和深度神经网络,构建船舶电气设备状态识别网络模型,结合电气设备的实时运行数据,识别船舶电气设备状态。测试结果显示,该方法能够确定数据中的连续异常数据和独立异常数据,并且完成所有异常数据的修正,保证数据的完整性;离散度结果均在0.016以下;能够完成电气设备正常状态、异常状态以及紧急状态的识别,最小均方根误差值均在0.0044以下,识别效果良好。

时频域多尺度交叉注意力融合的时间序列分类方法

针对时间序列子序列间的潜在信息交互不足导致分类准确率低的问题,提出时频域多尺度交叉注意力融合的时间序列分类方法TFFormer(Time-Frequency Transformer)。首先,将原始时间序列的时频域谱分别划分为等长子序列,经线性投影后加入位置信息解决时间序列的点值耦合问题;其次,通过改进的多头自注意力(IMHA)模块使模型关注更重要的序列特征,解决长时间序列的前后依赖问题;最后,构造多尺度时频域交叉注意力(CMA)模块增强时间序列在时域和频域之间的信息交互,使模型进一步挖掘序列的频域信息。实验结果表明,在Trace、StarLightCurves和UWaveGestureLibraryAll数据集上,相较于全卷积网络(FCN),所提方法的分类准确率分别提高了0.3、0.9和1.4个百分点,验证了通过增强时间序列时域和频域间的信息交互,可以提高模型收敛速度和分类精度。

基于细节增强的双分支实时语义分割网络

实时语义分割方法常利用双分支结构分别保存图像的浅层空间信息和深层语义信息。然而,当前基于双分支结构的实时语义分割方法重点研究语义特征的挖掘,忽略了空间特征的保持,导致网络无法精准地捕捉图像内物体的边界和纹理等细节特征,最终分割效果欠佳。针对以上问题,提出基于细节增强的双分支实时语义分割网络(DEDBNet),多阶段增强空间细节信息。首先,提出细节增强双向交互(DEBIM)模块,在分支间的交互阶段使用轻量空间注意力机制增强高分辨率特征图对细节信息的表达能力,促进空间细节特征在高低两分支上的流动,以加强网络对细节信息的学习能力;其次,设计局部细节注意力特征融合模块(LDAFF),在两分支末端特征融合的过程中同时建模全局语义信息和局部空间信息,解决不同层次特征图之间细节不连续的问题;此外,引入边界损失,在不影响模型速度的情况下引导网络浅层学习物体边界信息。所提网络在Cityscapes验证集上以92.3 frame/s的帧速率(FPS)获得78.2%的平均交并比(mIoU),在CamVid测试集上以202.8 frame/s获得79.2%的mIoU;与深度双分辨率网络(DDRNet-23-slim)相比,mIoU分别提高了1.1和4.5个百分点。实验结果表明,DEDBNet能够准确地分割场景图像,且满足实时性要求。

四川地区精细化降水预报融合订正试验及检验

精细化定量降水网格预报是天气预报业务的重点和难点,基于西南区域智能数值网格模式预报系统(Southwest China WRF-based Intelligent Numeric Grid forecast System,SWC-WINGS)1 km×1 km分辨率的小时降水预报,利用时间滞后和概率匹配方法开展融合订正试验,再利用中国气象局多源融合降水系统(CMA Multi-source Precipitation Analysis System,CMPAS)三源融合降水实况格点数据,对2022年7—8月四川地区的小时降水预报融合订正结果进行检验,并在四川盆地西部一次短时强降水天气过程中进行应用,结果表明:(1)时间滞后集合降水预报相较于模式降水预报,存在小量级预报过度,大量级预报过于保守的问题;(2)时间滞后结合概率匹配的降水预报融合订正方法有效提升了各量级降水预报的TS评分,尤其1~2 h预报时效提升显著,小时雨量超过0.1 mm、5 mm、10 mm和20 mm量级的TS评分平均提升率分别为7.2%、17.2%、28.3%和36.3%;(3)一次短时强降水天气过程的应用结果表明,时间滞后结合概率匹配的融合订正方法对模式小时降水预报有较好的改进效果,尤其对大量级降水预报有较强的订正能力。

基于面向TTE的新型数据综合系统的设计与实现

针对传统总线无法实时网络传输,新型总线时间触发以太网(TTE)兼容IP协议和IEEE802.3协议,但TT业务和BE业务需设备满足TTE协议,实际工程缺少适用性的问题,设计了一种基于FPGA的时间触发以太网数据综合系统。该数据综合系统以FPGA作为核心控制器,千兆网作为主要通信接口,在满足高速率、高宽带、高灵活的前提下,设计出数据综合模块,将多种传统总线与TTE总线良好兼容。试验结果表明:系统在同步精度高达40 ns、通信抖动仅为392 ns的前提下,数据转换收发无误码现象。同时将多种数据混合编帧,具备极高的稳定性与灵活性,为TTE网络数据服务于实际工程提供了可靠的解决方案。

基于注意力机制及多分支特征融合的实时语义分割算法

为了合理平衡语义分割中的精确度与实时性,基于快速卷积神经网络模型(Fast-SCNN)提出了一种基于注意力机制及多分支特征融合的实时语义分割算法模型。该算法模型首先通过注意力模块捕获空间特征之间的相互联系,增强空间细节信息;然后合理设计融合模块,最大化利用各分支信息,实现深层特征与浅层特征更好的融合;最后引入自适应特征增强注意力模块,捕获长距离像素间的相互依赖关系。实验结果表明,文中算法模型在Cityscapes数据集上获得了71.55%的分割精度,推理速度FPS达到97.6帧/s,模型参数量为1.39 M,验证了该算法所构成网络模型的有效性。

基于多特征融合时差网络的带式输送机区域违规行为识别

现有的煤矿井下带式输送机区域违规行为(如攀爬、跨越、倚靠带式输送机等)识别方法对特征提取不充分、难以考虑到行为时间差异,导致违规行为识别准确率不高。针对该问题,基于ResNet50模型,提出了一种基于多特征融合时差网络(MFFTDN)的带式输送机区域违规行为识别方法,将多特征融合和时间差分进行结合,对不同时间段的行为进行多特征融合。首先在原始模型ResNet50的第2和第3阶段引入短期多特征融合(STMFF)模块,将来自多个连续帧的时间和特征拼接在一起,再对融合后的特征进行时间差分计算,即相邻帧的特征差值,以在短期内捕捉局部动作变化。然后在原始模型ResNet50的第4阶段引入长期多特征融合(LTMFF)模块,将来自连续帧的短期多特征拼接在一起,再对相邻时间点的特征进行时间差分计算,以获取行为的长期多特征。最后将融合后的特征进行分类,输出识别结果。实验结果表明:①该方法的平均精度和准确率较原始模型ResNet50分别提高了8.18%和8.47%,说明同时引入STMFF和LTMFF模块能够有效提取到不同时间段的多特征信息。②该方法在自建煤矿井下带式输送机区域违规行为数据集上的准确率为89.62%,平均精度为89.30%,模型的参数量为197.2×10^(6)。③Grad−CAM热力图显示,该方法能够更有效地关注到违规行为的关键区域,精确捕捉到井下带式输送机区域的违规行为。

表面肌电与三轴信息融合的运动判断实验

为了提高基于表面肌电与三轴加速度信号的运动识别准确率,提出了一套多源信息融合处理的实验流程与方法。该方法利用5层离散小波变换对表面肌电信号进行分解,充分提取不同运动产生的肌电信号中各频域的特征信息;再将分解后的表面肌电信号与三轴加速度信号通过滑动窗口的方法进行特征融合,构造融合肌电与空间运动特征的特征图;最后用融合特征图对深度学习模型进行训练,并结合自动状态机进行最终运动状态的识别。实验结果表明,多源信息融合处理方法可以提高运动识别的准确性,总体识别精度分别达到了95.4%和89.2%。该方法在实时性与准确性上均有良好表现。