Defect Detection Model Using Time Series Data Augmentation and Transformation

Time-series data provide important information in many fields,and their processing and analysis have been the focus of much research.However,detecting anomalies is very difficult due to data imbalance,temporal dependence,and noise.Therefore,methodologies for data augmentation and conversion of time series data into images for analysis have been studied.This paper proposes a fault detection model that uses time series data augmentation and transformation to address the problems of data imbalance,temporal dependence,and robustness to noise.The method of data augmentation is set as the addition of noise.It involves adding Gaussian noise,with the noise level set to 0.002,to maximize the generalization performance of the model.In addition,we use the Markov Transition Field(MTF)method to effectively visualize the dynamic transitions of the data while converting the time series data into images.It enables the identification of patterns in time series data and assists in capturing the sequential dependencies of the data.For anomaly detection,the PatchCore model is applied to show excellent performance,and the detected anomaly areas are represented as heat maps.It allows for the detection of anomalies,and by applying an anomaly map to the original image,it is possible to capture the areas where anomalies occur.The performance evaluation shows that both F1-score and Accuracy are high when time series data is converted to images.Additionally,when processed as images rather than as time series data,there was a significant reduction in both the size of the data and the training time.The proposed method can provide an important springboard for research in the field of anomaly detection using time series data.Besides,it helps solve problems such as analyzing complex patterns in data lightweight.

Digital Transformation of Enterprise Human Resource Management Enabled by Big Data

With the continuous development of big data technology,the digital transformation of enterprise human resource management has become a development trend.Human resources is one of the most important resources of enterprises,which is crucial to the competitiveness of enterprises.Enterprises need to attract,retain,and motivate excellent employees,thereby enhancing the innovation ability of enterprises and improving competitiveness and market share in the market.To maintain advantages in the fierce market competition,enterprises need to adopt more scientific and effective human resource management methods to enhance organizational efficiency and competitiveness.At the same time,this paper analyzes the dilemma faced by enterprise human resource management,points out the new characteristics of enterprise human resource management enabled by big data,and puts forward feasible suggestions for enterprise digital transformation.

基于生成对抗Transformer的电力负荷数据异常检测

电力负荷异常数据将给电力系统规划、负荷预测以及用能分析等带来较大的负面影响,因此亟须对负荷数据异常进行检测与识别。首先,针对电力负荷数据异常分类、原因及其特征开展分析。其次,改进传统Transformer编码器结构,采用多头注意力层代替掩码多头注意力层,同时移除前馈网络,以提高模型对负荷时序序列的全局注意力。基于生成对抗网络(generative adversarial networks,GAN)生成器与判别器的博弈结构,提出一种改进的GAN-Transformer模型,以更好地捕捉趋势性特征并加速模型收敛。然后,引入多阶段映射与训练方法,综合焦点分数打分机制,通过分阶段负荷序列重构帮助模型更好地提取负荷数据异常特征。最后,算例分析结果表明,GAN-Transformer模型在负荷数据异常检测精确率、召回率、F_(1)值以及训练时间方面均具有更优的性能,验证了所提方法的有效性和优越性。文中研究工作为基于深度学习进一步实现电力负荷数据异常分类与数据修复提供了有益参考。

CNN-Transformer结合对比学习的高光谱与LiDAR数据协同分类

针对高光谱图像(hyperspectral images,HSI)与LiDAR数据多模态分类任务中的跨模态信息表达和特征对齐等问题,提出一种基于对比学习CNN-Transformer高光谱和LiDAR数据协同分类网络(Contrastive Learning based CNNTransformer Network,CLCT-Net)。CLCT-Net通过由ConvNeXt V2 Block构成的共有特征提取模块,获得不同模态间的共性特征,解决异构传感器数据之间语义对齐的问题。构建了包含空间-通道分支和光谱上下文分支的双分支HSI编码器,以及结合频域自注意力机制的LiDAR编码器,以获取更丰富的特征表示。利用集成对比学习进行分类,进一步提升多模态数据协同分类的精度。在Houston 2013和Trento数据集上的实验结果表明,相较于其他高光谱图像和Li‐DAR数据分类模型,本文所提模型获得了更高的地物分类精度,分别达到了92.01%和98.90%,实现了跨模态数据特征的深度挖掘和协同提取。

基于Transformer的司法文书命名实体识别方法

命名实体识别是自然语言处理领域的关键任务之一,是实现下游任务的基础。目前针对司法领域的相关研究相对较少,司法系统的信息化和智能化转型仍有许多问题亟需解决。相比其他领域的文本,司法文书存在专业性强、语料资源少等局限,导致现有的司法文书识别结果较低。因此,从以下3方面开展研究:首先,提出了一种多标签层级迭代的文本标注方式,可以对原始司法文书文本进行自动化标注,同时有效地提升司法文书命名实体识别任务的实体识别效果;其次,提出了一种交融式的Transformer神经网络模型,对汉字固有属性的深层特征进行了充分利用,用于对司法文书进行命名实体识别;最后,对所提出的标注方法和模型与其他神经网络模型进行了对比实验。所提出的文本标注方式可以较为准确地实现司法文书的标注任务;同时,所提出的模型在通用数据集中相对于对照模型有较大的提高,并在司法领域数据集中取得了良好的效果。

融合Transformer和卷积LSTM的轨迹分类网络

为了减少原始轨迹数据的噪声,充分提取轨迹的时空特征,提高基于轨迹数据的交通模式分类精度,提出一种融合堆叠降噪自编码器、Transformer和卷积长短期记忆网络的轨迹分类网络(networks fusing stacked denoising auto-encoder, Transformer and ConvLSTM,SDAETC)。通过堆叠降噪自编码器减少原始轨迹数据中的噪声;利用结合了Transformer的递归图自编码器,提取到更为丰富的时间特征,同时利用特征图自编码器提取空间特征;改进卷积长短期记忆网络,充分提取轨迹中的时空特征,并与提取到的时间特征和空间特征相融合,从而实现交通模式分类。实验结果表明,提出的SDAETC与基线模型相比,在GeoLife和SHL数据集上的准确率分别提升了1.8%和2%。此外,消融实验结果和模型训练时间分析表明,引入堆叠降噪自编码器、Transfomer和ConvLSTM虽然增加了时间消耗,但是对分类精度有积极贡献。

Logformer: Cascaded Transformer for System Log Anomaly Detection

Modern large-scale enterprise systems produce large volumes of logs that record detailed system runtime status and key events at key points.These logs are valuable for analyzing performance issues and understanding the status of the system.Anomaly detection plays an important role in service management and system maintenance,and guarantees the reliability and security of online systems.Logs are universal semi-structured data,which causes difficulties for traditional manual detection and pattern-matching algorithms.While some deep learning algorithms utilize neural networks to detect anomalies,these approaches have an over-reliance on manually designed features,resulting in the effectiveness of anomaly detection depending on the quality of the features.At the same time,the aforementioned methods ignore the underlying contextual information present in adjacent log entries.We propose a novel model called Logformer with two cascaded transformer-based heads to capture latent contextual information from adjacent log entries,and leverage pre-trained embeddings based on logs to improve the representation of the embedding space.The proposed model achieves comparable results on HDFS and BGL datasets in terms of metric accuracy,recall and F1-score.Moreover,the consistent rise in F1-score proves that the representation of the embedding spacewith pre-trained embeddings is closer to the semantic information of the log.

Data Transformation for Super-totalstation Positioning System Integrated by GPS and Totalstation

This paper focuses on the integration and data transformation between GPS and totalstation.It emphasizes on the way to transfer the WGS84 Cartesian coordinates to the local two_dimensional plane coordinates and the orthometric height GPS receiver,totalstation,radio,notebook computer and the corresponding software work together to form a new surveying system,the super_totalstation positioning system(SPS) and a new surveying model for terrestrial surveying.With the help of this system,the positions of detail points can be measured.

A Recursive High Payload Reversible Data Hiding Using Integer Wavelet and Arnold Transform

Reversible data hiding is an information hiding technique that requires the retrieval of the error free cover image after the extraction of the secret image.We suggested a technique in this research that uses a recursive embedding method to increase capacity substantially using the Integer wavelet transform and the Arnold transform.The notion of Integer wavelet transforms is to ensure that all coefficients of the cover images are used during embedding with an increase in payload.By scrambling the cover image,Arnold transform adds security to the information that gets embedded and also allows embedding more information in each iteration.The hybrid combination of Integer wavelet transform and Arnold transform results to build a more efficient and secure system.The proposed method employs a set of keys to ensure that information cannot be decoded by an attacker.The experimental results show that it aids in the development of a more secure storage system and withstand few tampering attacks The suggested technique is tested on many image formats,including medical images.Various performance metrics proves that the retrieved cover image and hidden image are both intact.This System is proven to withstand rotation attack as well.

Fault Diagnosis of Power Transformer Based on Improved ACGAN Under Imbalanced Data

The imbalance of dissolved gas analysis(DGA)data will lead to over-fitting,weak generalization and poor recognition performance for fault diagnosis models based on deep learning.To handle this problem,a novel transformer fault diagnosis method based on improved auxiliary classifier generative adversarial network(ACGAN)under imbalanced data is proposed in this paper,which meets both the requirements of balancing DGA data and supplying accurate diagnosis results.The generator combines one-dimensional convolutional neural networks(1D-CNN)and long short-term memories(LSTM),which can deeply extract the features from DGA samples and be greatly beneficial to ACGAN’s data balancing and fault diagnosis.The discriminator adopts multilayer perceptron networks(MLP),which prevents the discriminator from losing important features of DGA data when the network is too complex and the number of layers is too large.The experimental results suggest that the presented approach can effectively improve the adverse effects of DGA data imbalance on the deep learning models,enhance fault diagnosis performance and supply desirable diagnosis accuracy up to 99.46%.Furthermore,the comparison results indicate the fault diagnosis performance of the proposed approach is superior to that of other conventional methods.Therefore,the method presented in this study has excellent and reliable fault diagnosis performance for various unbalanced datasets.In addition,the proposed approach can also solve the problems of insufficient and imbalanced fault data in other practical application fields.

基于Vision Transformer的小麦病害图像识别算法

小麦白粉病、赤霉病和锈病是危害小麦产量的三大病害。为提高小麦病害图像的识别准确率,构建一种基于Vision Transformer的小麦病害图像识别算法。首先,通过田间拍摄的方式收集包含小麦白粉病、赤霉病和锈病3种病害在内的小麦病害图像,并对原始图像进行预处理,建立小麦病害图像识别数据集;然后,基于改进的Vision Transformer构建小麦病害图像识别算法,分析不同迁移学习方式和数据增强对模型识别效果的影响。试验可知,全参数迁移学习和数据增强能明显提高Vision Transformer模型的收敛速度和识别精度。最后,在相同时间条件下,对比Vision Transformer、AlexNet和VGG16算法在相同数据集上的表现。试验结果表明,Vision Transformer模型对3种小麦病害图像的平均识别准确率为96.81%,相较于AlexNet和VGG16模型识别准确率分别提高6.68%和4.94%。

基于时空Transformer的多空间尺度交通预测模型

准确的交通预测对提高智能交通系统的效率至关重要。交通系统的空间依赖不仅体现在道路的相连关系上,更重要的是由道路属性、区域功能等因素形成的隐藏空间依赖。另外,交通数据之间的时间依赖具有严格的相对位置关系,忽略这一问题将难以实现准确的交通预测。为了解决这些问题,提出了一种基于时空Transformer的多空间尺度交通预测模型(MSS-STT)。MSS-STT使用多个特定的Transformer网络对不同的空间尺度建模,以捕捉隐藏空间依赖,同时使用图卷积网络来学习静态空间特征。接着,使用门控机制将不同空间尺度的空间依赖与静态空间特征根据各自对预测的重要性进行融合。最后,根据时间序列中不同相对位置对预测的不同贡献来提取不同的时间依赖关系。在PeMS数据集上的实验结果表明,MSS-STT优于最先进的基线。

基于变量选择与Transformer模型的中长期电力负荷预测方法

准确且有效的负荷预测对于电力系统的实时运行和调度非常重要。提出了一种融合变量选择与稀疏Transformer模型的预测方法,将静态变量和时序变量作为输入,充分发挥静态变量在全局时间范围内的信息增强作用,基于门控机制设计变量分权组件,根据变量与预测结果的相关性,赋予变量不同的权重。设计了双层编码结构,进行时序特征提取,对注意力进行稀疏处理,通过多变量输入对未来时刻负荷进行预测。基于真实电力负荷数据的实验表明,本文模型能够提高中长期负荷预测精度和效率。

基于Transformer的城市三角网格语义分割方法

对城市三角网格(Urban Triangle Mesh)数据进行语义分割以识别不同类别的物体,是理解和分析三维城市场景的一种非常重要的方法.城市三角网格是一种具有丰富空间拓扑关系的三维空间几何数据,包含大量的几何信息,然而,现有的方法仅仅单独对每种几何信息进行特征提取,然后简单地融合再进行语义分割,难以利用几何信息之间的关联性,对个别物体的分割性能不佳.为了解决上述问题,提出一种基于自注意力机制Transformer的模型UMeT(Urban Mesh Transformer),其由多层感知机和MeshiT(Mesh in Transformer)模块构成,不仅可以利用多层感知机提取高维特征,还可以利用MeshiT模块计算各种几何信息之间的关联性,有效挖掘城市三角网格数据中隐含的关联.实验证明,UMeT能提取高维特征,同时保证城市三角网格数据的空间不变性,从而提升了语义分割的准确性.

基于CWT和优化Swin Transformer的风电齿轮箱故障诊断方法

针对传统故障诊断方法在风电齿轮箱运行故障诊断应用上的不足,提出一种基于小波变换(continuous wavelet transform, CWT)和优化Swin Transformer的风电齿轮箱故障诊断方法。该方法利用小波变换将风电齿轮箱振动信号转换为时频图;使用SuperMix数据增强算法对样本进行扩充;利用迁移学习技术将模型预训练参数用于训练和优化Swin Transformer模型;将训练完成的优化Swin Transformer模型应用于风场实际运维数据进行对比验证,分类准确率达到99.67%。验证结果表明该方法能够有效地实现风电齿轮箱故障诊断,并提高模型的识别准确率。

基于CNN和Transformer点云图像融合的道路检测

针对道路检测模型易受光线及阴影影响而导致精度不高及道路边缘分割不准确的问题,提出一种基于Transformer和卷积神经网络模型混合且以RGB图像和三维激光雷达点云共同为输入的道路分割算法,实现了无人车在自动驾驶过程中对所在行驶道路的精确感知。在KITTI道路数据集上的实验结果表明:与现有的道路检测模型相比,本文方法在分割精度方面具有较好的性能。

基于Transformer神经网络的锂电池热失控多数据融合探测

为满足对锂离子电池热失控高效准确探测的需求,设计了一种锂电池热失控试验平台,并利用STM32F103ZET6单片机连接了CO、CO_(2)、H2和热敏电阻NTC共4种传感器,实时采集特征参量。同时,利用PyroSim模拟试验环境,生成高质量的模拟数据,以补充试验数据。基于PyTorch平台,设计了一个Transformer神经网络,能够输出锂电池的正常、预警和热失控3种状态。通过使用试验数据和模拟数据进行训练,实现了对锂电池热失控的融合探测,相比于其他算法有一定的优势。

Performance of Continuous Wavelet Transform over Fourier Transform in Features Resolutions

This study presents a comparative analysis of two image enhancement techniques, Continuous Wavelet Transform (CWT) and Fast Fourier Transform (FFT), in the context of improving the clarity of high-quality 3D seismic data obtained from the Tano Basin in West Africa, Ghana. The research focuses on a comparative analysis of image clarity in seismic attribute analysis to facilitate the identification of reservoir features within the subsurface structures. The findings of the study indicate that CWT has a significant advantage over FFT in terms of image quality and identifying subsurface structures. The results demonstrate the superior performance of CWT in providing a better representation, making it more effective for seismic attribute analysis. The study highlights the importance of choosing the appropriate image enhancement technique based on the specific application needs and the broader context of the study. While CWT provides high-quality images and superior performance in identifying subsurface structures, the selection between these methods should be made judiciously, taking into account the objectives of the study and the characteristics of the signals being analyzed. The research provides valuable insights into the decision-making process for selecting image enhancement techniques in seismic data analysis, helping researchers and practitioners make informed choices that cater to the unique requirements of their studies. Ultimately, this study contributes to the advancement of the field of subsurface imaging and geological feature identification.

基于Transformer网络特征融合的色纺织物颜色表征模型

针对色纺织物特有的呈色结构以及常用单一颜色测量工具导致的数据特征局限的问题,基于Transformer网络建立多源异构数据融合颜色表征模型。从分光计数据与图像数据中分别提取光谱特征与纹理特征,使用Transformer网络融合多源异构数据特征,充分利用其互补特性,有效完备地表征色纺织物颜色信息。结果表明,本文构建的颜色表征模型能够对在较大变化范围内的染色纤维质量配比差异和由于染色纤维表面不均匀分布的颜色变化进行有效表征;在分光光度仪6、10和25 mm测量孔径下,融合特征差异与纤维配比差异的相关系数均高于85%;相较于单一光谱特征和单一图像特征,本文方法的相关系数提升10%以上,具有理想的鲁棒性。

基于随机增强Swin-Tiny Transformer的玉米病害识别及应用

针对图像识别中获取全局特征的局限性及难以提升识别准确性的问题,提出一种基于随机增强Swin-Tiny Transformer轻量级模型的图像识别方法.该方法在预处理阶段结合基于随机数据增强(random data augmentation based enhancement,RDABE)算法对图像特征进行增强,并采用Transformer的自注意力机制,以获得更全面的高层视觉语义信息.通过在玉米病害数据集上优化Swin-Tiny Transformer模型并进行参数微调,在农业领域的玉米病害上验证了该算法的适用性,实现了更精确的病害检测.实验结果表明,基于随机增强的轻量级Swin-Tiny+RDABE模型对玉米病害图像识别准确率达93.5867%.在参数权重一致,与性能优秀的轻量级Transformer、卷积神经网络(CNN)系列模型对比的实验结果表明,改进的模型准确率比Swin-Tiny Transformer,Deit3_Small,Vit_Small,Mobilenet_V3_Small,ShufflenetV2和Efficientnet_B1_Pruned模型提高了1.1877%~4.9881%,且能迅速收敛.