A multiscale adaptive framework based on convolutional neural network:Application to fluid catalytic cracking product yield prediction

Since chemical processes are highly non-linear and multiscale,it is vital to deeply mine the multiscale coupling relationships embedded in the massive process data for the prediction and anomaly tracing of crucial process parameters and production indicators.While the integrated method of adaptive signal decomposition combined with time series models could effectively predict process variables,it does have limitations in capturing the high-frequency detail of the operation state when applied to complex chemical processes.In light of this,a novel Multiscale Multi-radius Multi-step Convolutional Neural Network(Msrt Net)is proposed for mining spatiotemporal multiscale information.First,the industrial data from the Fluid Catalytic Cracking(FCC)process decomposition using Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN)extract the multi-energy scale information of the feature subset.Then,convolution kernels with varying stride and padding structures are established to decouple the long-period operation process information encapsulated within the multi-energy scale data.Finally,a reconciliation network is trained to reconstruct the multiscale prediction results and obtain the final output.Msrt Net is initially assessed for its capability to untangle the spatiotemporal multiscale relationships among variables in the Tennessee Eastman Process(TEP).Subsequently,the performance of Msrt Net is evaluated in predicting product yield for a 2.80×10^(6) t/a FCC unit,taking diesel and gasoline yield as examples.In conclusion,Msrt Net can decouple and effectively extract spatiotemporal multiscale information from chemical process data and achieve a approximately reduction of 30%in prediction error compared to other time-series models.Furthermore,its robustness and transferability underscore its promising potential for broader applications.

Robust Damage Detection and Localization Under Complex Environmental Conditions Using Singular Value Decomposition-based Feature Extraction and One-dimensional Convolutional Neural Network

Ultrasonic guided wave is an attractive monitoring technique for large-scale structures but is vulnerable to changes in environmental and operational conditions(EOC),which are inevitable in the normal inspection of civil and mechanical structures.This paper thus presents a robust guided wave-based method for damage detection and localization under complex environmental conditions by singular value decomposition-based feature extraction and one-dimensional convolutional neural network(1D-CNN).After singular value decomposition-based feature extraction processing,a temporal robust damage index(TRDI)is extracted,and the effect of EOCs is well removed.Hence,even for the signals with a very large temperature-varying range and low signal-to-noise ratios(SNRs),the final damage detection and localization accuracy retain perfect 100%.Verifications are conducted on two different experimental datasets.The first dataset consists of guided wave signals collected from a thin aluminum plate with artificial noises,and the second is a publicly available experimental dataset of guided wave signals acquired on a composite plate with a temperature ranging from 20℃to 60℃.It is demonstrated that the proposed method can detect and localize the damage accurately and rapidly,showing great potential for application in complex and unknown EOC.

Remaining Useful Life Prediction of Aeroengine Based on Principal Component Analysis and One-Dimensional Convolutional Neural Network

In order to directly construct the mapping between multiple state parameters and remaining useful life(RUL),and reduce the interference of random error on prediction accuracy,a RUL prediction model of aeroengine based on principal component analysis(PCA)and one-dimensional convolution neural network(1D-CNN)is proposed in this paper.Firstly,multiple state parameters corresponding to massive cycles of aeroengine are collected and brought into PCA for dimensionality reduction,and principal components are extracted for further time series prediction.Secondly,the 1D-CNN model is constructed to directly study the mapping between principal components and RUL.Multiple convolution and pooling operations are applied for deep feature extraction,and the end-to-end RUL prediction of aeroengine can be realized.Experimental results show that the most effective principal component from the multiple state parameters can be obtained by PCA,and the long time series of multiple state parameters can be directly mapped to RUL by 1D-CNN,so as to improve the efficiency and accuracy of RUL prediction.Compared with other traditional models,the proposed method also has lower prediction error and better robustness.

Establishing and validating a spotted tongue recognition and extraction model based on multiscale convolutional neural network

Objective In tongue diagnosis,the location,color,and distribution of spots can be used to speculate on the viscera and severity of the heat evil.This work focuses on the image analysis method of artificial intelligence(AI)to study the spotted tongue recognition of traditional Chinese medicine(TCM).Methods A model of spotted tongue recognition and extraction is designed,which is based on the principle of image deep learning and instance segmentation.This model includes multiscale feature map generation,region proposal searching,and target region recognition.Firstly,deep convolution network is used to build multiscale low-and high-abstraction feature maps after which,target candidate box generation algorithm and selection strategy are used to select high-quality target candidate regions.Finally,classification network is used for classifying target regions and calculating target region pixels.As a result,the region segmentation of spotted tongue is obtained.Under non-standard illumination conditions,various tongue images were taken by mobile phones,and experiments were conducted.Results The spotted tongue recognition achieved an area under curve(AUC)of 92.40%,an accuracy of 84.30%with a sensitivity of 88.20%,a specificity of 94.19%,a recall of 88.20%,a regional pixel accuracy index pixel accuracy(PA)of 73.00%,a mean pixel accuracy(m PA)of73.00%,an intersection over union(Io U)of 60.00%,and a mean intersection over union(mIo U)of 56.00%.Conclusion The results of the study verify that the model is suitable for the application of the TCM tongue diagnosis system.Spotted tongue recognition via multiscale convolutional neural network(CNN)would help to improve spot classification and the accurate extraction of pixels of spot area as well as provide a practical method for intelligent tongue diagnosis of TCM.

Infrasound Event Classification Fusion Model Based on Multiscale SE-CNN and BiLSTM

The classification of infrasound events has considerable importance in improving the capability to identify the types of natural disasters.The traditional infrasound classification mainly relies on machine learning algorithms after artificial feature extraction.However,guaranteeing the effectiveness of the extracted features is difficult.The current trend focuses on using a convolution neural network to automatically extract features for classification.This method can be used to extract signal spatial features automatically through a convolution kernel;however,infrasound signals contain not only spatial information but also temporal information when used as a time series.These extracted temporal features are also crucial.If only a convolution neural network is used,then the time dependence of the infrasound sequence will be missed.Using long short-term memory networks can compensate for the missing time-series features but induces spatial feature information loss of the infrasound signal.A multiscale squeeze excitation–convolution neural network–bidirectional long short-term memory network infrasound event classification fusion model is proposed in this study to address these problems.This model automatically extracted temporal and spatial features,adaptively selected features,and also realized the fusion of the two types of features.Experimental results showed that the classification accuracy of the model was more than 98%,thus verifying the effectiveness and superiority of the proposed model.

Grid Side Distributed Energy Storage Cloud Group End Region Hierarchical Time-Sharing Configuration Algorithm Based onMulti-Scale and Multi Feature Convolution Neural Network

There is instability in the distributed energy storage cloud group end region on the power grid side.In order to avoid large-scale fluctuating charging and discharging in the power grid environment and make the capacitor components showa continuous and stable charging and discharging state,a hierarchical time-sharing configuration algorithm of distributed energy storage cloud group end region on the power grid side based on multi-scale and multi feature convolution neural network is proposed.Firstly,a voltage stability analysis model based onmulti-scale and multi feature convolution neural network is constructed,and the multi-scale and multi feature convolution neural network is optimized based on Self-OrganizingMaps(SOM)algorithm to analyze the voltage stability of the cloud group end region of distributed energy storage on the grid side under the framework of credibility.According to the optimal scheduling objectives and network size,the distributed robust optimal configuration control model is solved under the framework of coordinated optimal scheduling at multiple time scales;Finally,the time series characteristics of regional power grid load and distributed generation are analyzed.According to the regional hierarchical time-sharing configuration model of“cloud”,“group”and“end”layer,the grid side distributed energy storage cloud group end regional hierarchical time-sharing configuration algorithm is realized.The experimental results show that after applying this algorithm,the best grid side distributed energy storage configuration scheme can be determined,and the stability of grid side distributed energy storage cloud group end region layered timesharing configuration can be improved.

Convolutional Neural Network Based on Spatial Pyramid for Image Classification

A novel convolutional neural network based on spatial pyramid for image classification is proposed.The network exploits image features with spatial pyramid representation.First,it extracts global features from an original image,and then different layers of grids are utilized to extract feature maps from different convolutional layers.Inspired by the spatial pyramid,the new network contains two parts,one of which is just like a standard convolutional neural network,composing of alternating convolutions and subsampling layers.But those convolution layers would be averagely pooled by the grid way to obtain feature maps,and then concatenated into a feature vector individually.Finally,those vectors are sequentially concatenated into a total feature vector as the last feature to the fully connection layer.This generated feature vector derives benefits from the classic and previous convolution layer,while the size of the grid adjusting the weight of the feature maps improves the recognition efficiency of the network.Experimental results demonstrate that this model improves the accuracy and applicability compared with the traditional model.

Feature Fusion Multi_XMNet Convolution Neural Network for Clothing Image Classification

Faced with the massive amount of online shopping clothing images,how to classify them quickly and accurately is a challenging task in image classification.In this paper,we propose a novel method,named Multi_XMNet,to solve the clothing images classification problem.The proposed method mainly consists of two convolution neural network(CNN)branches.One branch extracts multiscale features from the whole expressional image by Multi_X which is designed by improving the Xception network,while the other extracts attention mechanism features from the whole expressional image by MobileNetV3-small network.Both multiscale and attention mechanism features are aggregated before making classification.Additionally,in the training stage,global average pooling(GAP),convolutional layers,and softmax classifiers are used instead of the fully connected layer to classify the final features,which speed up model training and alleviate the problem of overfitting caused by too many parameters.Experimental comparisons are made in the public DeepFashion dataset.The experimental results show that the classification accuracy of this method is 95.38%,which is better than InceptionV3,Xception and InceptionV3_Xception by 5.58%,3.32%,and 2.22%,respectively.The proposed Multi_XMNet image classification model can help enterprises and researchers in the field of clothing e-commerce to automaticly,efficiently and accurately classify massive clothing images.

Fault Line Detection Using Waveform Fusion and One-dimensional Convolutional Neural Network in Resonant Grounding Distribution Systems

Effective features are essential for fault diagnosis.Due to the faint characteristics of a single line-to-ground(SLG)fault,fault line detection has become a challenge in resonant grounding distribution systems.This paper proposes a novel fault line detection method using waveform fusion and one-dimensional convolutional neural networks(1-D CNN).After an SLG fault occurs,the first-half waves of zero-sequence currents are collected and superimposed with each other to achieve waveform fusion.The compelling feature of fused waveforms is extracted by 1-D CNN to determine whether the fused waveform source contains the fault line.Then,the 1-D CNN output is used to update the value of the counter in order to identify the fault line.Given the lack of fault data in existing distribution systems,the proposed method only needs a small quantity of data for model training and fault line detection.In addition,the proposed method owns fault-tolerant performance.Even if a few samples are misjudged,the fault line can still be detected correctly based on the full output results of 1-D CNN.Experimental results verified that the proposed method can work effectively under various fault conditions.

Multidimensional attention and multiscale upsampling for semantic segmentation

Semantic segmentation is for pixel-level classification tasks,and contextual information has an important impact on the performance of segmentation.In order to capture richer contextual information,we adopt ResNet as the backbone network and designs an encoder-decoder architecture based on multidimensional attention(MDA)module and multiscale upsampling(MSU)module.The MDA module calculates the attention matrices of the three dimensions to capture the dependency of each position,and adaptively captures the image features.The MSU module adopts parallel branches to capture the multiscale features of the images,and multiscale feature aggregation can enhance contextual information.A series of experiments demonstrate the validity of the model on Cityscapes and Camvid datasets.

基于Transformer的多尺度遥感语义分割网络

为了提升遥感图像语义分割效果,本文针对分割目标类间方差小、类内方差大的特点,从全局上下文信息和多尺度语义特征2个关键点提出一种基于Transformer的多尺度遥感语义分割网络(muliti-scale Transformer network,MSTNet)。其由编码器和解码器2个部分组成,编码器包含基于Transformer改进的视觉注意网络(visual attention network,VAN)主干和基于空洞空间金字塔池化(atrous spatial pyramid pooling, ASPP)结构改进的多尺度语义特征提取模块(multi-scale semantic feature extraction module, MSFEM)。解码器采用轻量级多层感知器(multi-layer perception,MLP)配合编码器设计,充分分析所提取的包含全局上下文信息和多尺度表示的语义特征。MSTNet在2个高分辨率遥感语义分割数据集ISPRS Potsdam和LoveDA上进行验证,平均交并比(mIoU)分别达到79.50%和54.12%,平均F1-score(m F1)分别达到87.46%和69.34%,实验结果验证了本文所提方法有效提升了遥感图像语义分割的效果。

采用多尺度自适应选择卷积神经网络的轴承故障诊断研究

针对轴承故障诊断方法中传统多尺度卷积神经网络对不同尺度的特征只是简单拼接,而未考虑不同尺度的特征差异的问题,提出一种多尺度自适应选择卷积神经网络轴承故障诊断模型(MSASCNN)。通过不同大小的宽卷积筛选原始轴承振动信号中的特征,合并为初始特征;构建多尺度自适应选择卷积块,提取不同尺度的特征,利用改进的注意力机制自适应调整不同尺度的特征权重,加入残差连接,防止模型退化;通过分类器完成轴承故障诊断。在凯斯西储大学轴承数据集和XJTU-SY轴承数据集上的实验结果表明:在模型改进实验中,与没有改进注意力机制的模型相比,所提模型的轴承故障诊断准确率提升了1.98%;在不同信噪比的噪声干扰环境中,所提模型的轴承故障诊断准确率均高于93%。

摩尔纹图案自动去除技术综述

如今,数码相机和智能手机在人们的生活中扮演着越来越重要的角色,已经成为人们感知世界、记录信息和沟通交流的主要工具.当使用这些设备拍摄电子屏幕时,显示设备和摄像头传感器网格之间往往会发生混叠,通常导致图片中存在不规则分布的摩尔纹干扰图案,从而严重影响了拍摄图像的视觉质量效果.因此,摩尔纹图案去除方法研究对于拍摄图像的后期处理具有重要意义.为此,详细梳理了摩尔纹去除研究的发展脉络,并根据不同的适用场景和技术实现将现有方法分为2类:基于先验知识的摩尔纹去除方法和基于深度学习的摩尔纹去除方法.鉴于深度学习网络中训练数据集的收集和对齐方式不同,该类方法又分为基于卷积神经网络(CNN)的摩尔纹去除方法和基于生成式对抗网络(GAN)的摩尔纹去除方法.在此基础上,选择相同的公开数据集,对主流的深度学习方法进行算法实现和性能对比分析,并分别总结了各类方法的优缺点.最后,对未来的研究方向进行展望.

基于CBAM-CNN的涡旋压缩机故障诊断

针对涡旋压缩机振动信号不平稳和噪声情况下故障振动信号弱、需要人为提取故障特征以及准确率有待进一步提高的问题,提出基于多尺度注意力机制(convolutional block attention mechanism,简称CBAM)-卷积神经网络(convolutional neural network,简称CNN)涡旋压缩机故障诊断方法。首先,通过多个不同尺度的卷积核对振动信号转化为灰度图的故障特征进行全面提取,并引入注意力机制,通过调整权重值的方式提取重要的故障特征;其次,利用降维卷积模块、深度可分离卷积模块和残差模块提取更高维度的深层次故障特征,提升网络计算效率;最后,设置舍弃率为0.5的Dropout层防止过拟合,提升了网络的鲁棒性、抗干扰能力和泛化能力。实验结果证明,该方法在无噪声和添加不同信噪比噪声的情况下,均能有效地对涡旋压缩机故障进行分类,具有更高的识别准确性和更快的收敛能力。

融合注意力机制的刀具磨损预测方法

刀具的磨损状态影响着工件表面质量与加工稳定性,故实现其磨损量的准确监测对于保证加工可靠性、维持生产加工连续性具有积极作用.为进一步提高刀具磨损预测模型的泛化性能和准确度,提出一种融合注意力机制的多尺度卷积双向门控循环(multiscale convolutional bidirectional gated recurrent unit-attention,MSCBGRU-A)神经网络的刀具磨损预测方法,其由特征拓展模块、多尺度卷积模块、双向GRU模块、注意力模块、回归模块组成.首先,将切削力、声发射、振动信号作为输入信号,输入信号通过多尺度卷积模块获得多个尺度的刀具磨损输出特征图,将多个卷积通道输出的特征图输入到连接层进行首尾和层叠两种方式的连接来获得两种输出数据.然后,将两种输出数据分别输入到双向GRU模块与注意力模块,通过双向GRU模块学习输出特征图动态变化来获取时序特征,通过注意力模块对多尺度卷积神经网络的输出进行权值分配,强化对刀具磨损预测结果贡献度更大的特征.最后,通过回归模块对磨损值进行预测.经过对比实验引入混合域注意力机制的基于卷积块的注意力机制(convolutional block attention mechanism,CBAM),获得MSCBGRU-CBAM模型,并且通过绘制CBAM的注意力权重图证明注意力机制可以自适应地关注与刀具磨损更相关的特征.与其他深度学习模型进行对比实验表明,MSCBGRU-CBAM模型具有最高的预测精度,且与未使用注意力机制的MSCBGRU模型相比,RMSE降低19.3%,MAE降低17.7%,R 2提高2.7%.

比例融合与多层规模感知的人群计数方法

针对密集场景下人群图像拍摄视角或距离多变造成的多尺度特征获取不足、融合不佳和全局特征利用不充分等问题,提出一种比例融合与多层规模感知的人群计数网络。首先采用骨干网络VGG16提取人群密度初始特征;其次,设计多层规模感知模块,获得人群多尺度信息的丰富表达;再次,提出比例融合策略,根据卷积层捕获的特征权重重构多尺度信息,提取显著性人群特征;最后,采用卷积回归策略进行密度图的回归。同时,提出一种局部一致性损失函数,通过区域化密度图的方式增强生成密度图与真实密度图的相似度,提高计数性能。在多个人群数据集上的试验结果表明,所提模型优于近年人群计数的先进方法,且在车辆计数上有较好推广性。

基于轻量语义分割网络的遥感土地覆盖分类

高分辨率遥感图像有丰富的空间特征,针对遥感土地覆盖方法中模型复杂,边界模糊和多尺度分割等问题,提出了一种基于边界与多尺度信息的轻量化语义分割网络.首先,使用轻量化的MobileNetV3分类器,采用深度可分离卷积来减少计算量.其次,使用自顶向下和自底向上的特征金字塔结构来进行多尺度分割.接着,设计了一个边界增强模块,为分割任务提供丰富的边界细节信息.然后,设计了一个特征融合模块,融合边界与多尺度语义特征.最后,使用交叉熵损失函数和Dice损失函数来处理样本不平衡的问题.在WHDLD数据集的平均交并比达到了59.64%,总体精度达到了87.68%.在DeepGlobe数据集的平均交并比达到了70.42%,总体精度达到了88.81%.实验结果表明,该模型能快速有效地实现遥感图像土地覆盖分类.

基于注意力机制的多尺度融合人群计数算法

针对人群计数图像人头尺度变化大、背景噪声高等问题,提出一种基于注意力机制的多尺度融合人群计数算法,以充分聚合多尺度信息,并有效区分背景噪声。构建基于残差连接的空洞空间金字塔池化,通过残差结构以及多个不同扩张率的空洞卷积在捕获多尺度头部目标特征的同时融入浅层特征图的空间细节信息,提高特征图质量;构建跨层多尺度特征融合模块,融合浅层和深层分支不同大小的边缘细节信息和上下文语义信息,并设计基于多分支的特征融合模块,融合不同感受野大小的多尺度信息以缓解大规模人头尺度变化的问题;构建基于矩阵相似运算的通道和空间注意力机制模块提取像素级特征权重,加强网络对于背景和人头目标的判别能力,自适应矫正位置信息。实验结果表明,相比11种对比算法的最优值,所提算法在SHA数据集上的平均绝对误差和均方根误差指标降低1.4%、4.2%,在UCF_CC_50数据集上降低4.9%、1.8%,能够精确地预测人群分布状态和估计人群数量,生成高质量的人群密度图。

结合密集残差块和注意力的真实图像去噪网络

为有效去除真实图像噪声并保留图像边缘信息,提出一种结合密集网络思想和并行极化自注意力机制的真实去噪算法。使用3条并行结构处理不同尺度的特征信息,其中每条分支由两个密集注意力块串联而成,形成残差结构。使用选择性核融合机制,获取不同深度下的特征信息,将其融合并使用注意力机制去除冗余信息,获取干净图像。实验结果表明,该算法在SIDD、DND、PolyU测试集上的峰值信噪比分别为39.32 dB、39.52 dB和37.36 dB,结构相似性分别为0.908、0.951和0.952,在SIDD和PolyU测试集上的图像通用质量指标值为0.992和0.982,在去噪任务上可以达到较好的性能,提高了图像视觉的质量。

基于多尺度局部与全局上下文信息的钢材缺陷检测方法

钢材表面缺陷对许多工业产品的质量和性能有重大影响,会给生产带来巨大的经济损失。因此,对钢材表面进行实时监控,及时发现缺陷是非常有必要的。为提升对尺度差异较大、背景复杂的钢材表面缺陷的检测性能,提出一种基于多尺度局部与全局上下文信息的钢材缺陷检测模型。该模型使用卷积神经网络模型中带下采样机制的卷积操作,获取粗糙多尺度局部特征图,再使用自注意力机制分别在每一尺度作用于经过卷积提取的局部特征图,以获取像素间的长相关信息(如图像划痕、斑块、夹杂物等),增强缺陷类间判别能力;然后,采用特征金字塔结构,融合多尺度的特征图,以此提升对多尺度目标的检测能力;最后,引入通道与空间注意力模块与WIoU损失函数。实验结果表明,相比于Faster RCNN和EDDN等模型,该方法对于提升钢材表面缺陷检测性能行之有效。