Rockburst Intensity Grade Prediction Model Based on Batch Gradient Descent and Multi-Scale Residual Deep Neural Network

Rockburst is a phenomenon in which free surfaces are formed during excavation,which subsequently causes the sudden release of energy in the construction of mines and tunnels.Light rockburst only peels off rock slices without ejection,while severe rockburst causes casualties and property loss.The frequency and degree of rockburst damage increases with the excavation depth.Moreover,rockburst is the leading engineering geological hazard in the excavation process,and thus the prediction of its intensity grade is of great significance to the development of geotechnical engineering.Therefore,the prediction of rockburst intensity grade is one problem that needs to be solved urgently.By comprehensively considering the occurrence mechanism of rockburst,this paper selects the stress index(σθ/σc),brittleness index(σ_(c)/σ_(t)),and rock elastic energy index(Wet)as the rockburst evaluation indexes through the Spearman coefficient method.This overcomes the low accuracy problem of a single evaluation index prediction method.Following this,the BGD-MSR-DNN rockburst intensity grade prediction model based on batch gradient descent and a multi-scale residual deep neural network is proposed.The batch gradient descent(BGD)module is used to replace the gradient descent algorithm,which effectively improves the efficiency of the network and reduces the model training time.Moreover,the multi-scale residual(MSR)module solves the problem of network degradation when there are too many hidden layers of the deep neural network(DNN),thus improving the model prediction accuracy.The experimental results reveal the BGDMSR-DNN model accuracy to reach 97.1%,outperforming other comparable models.Finally,actual projects such as Qinling Tunnel and Daxiangling Tunnel,reached an accuracy of 100%.The model can be applied in mines and tunnel engineering to realize the accurate and rapid prediction of rockburst intensity grade.

Ash Detection of Coal Slime Flotation Tailings Based on Chromatographic Filter Paper Sampling and Multi-Scale Residual Network

The detection of ash content in coal slime flotation tailings using deep learning can be hindered by various factors such as foam,impurities,and changing lighting conditions that disrupt the collection of tailings images.To address this challenge,we present a method for ash content detection in coal slime flotation tailings.This method utilizes chromatographic filter paper sampling and a multi-scale residual network,which we refer to as MRCN.Initially,tailings are sampled using chromatographic filter paper to obtain static tailings images,effectively isolating interference factors at the flotation site.Subsequently,the MRCN,consisting of a multi-scale residual network,is employed to extract image features and compute ash content.Within the MRCN structure,tailings images undergo convolution operations through two parallel branches that utilize convolution kernels of different sizes,enabling the extraction of image features at various scales and capturing a more comprehensive representation of the ash content information.Furthermore,a channel attention mechanism is integrated to enhance the performance of the model.The combination of the multi-scale residual structure and the channel attention mechanism within MRCN results in robust capabilities for image feature extraction and ash content detection.Comparative experiments demonstrate that this proposed approach,based on chromatographic filter paper sampling and the multi-scale residual network,exhibits significantly superior performance in the detection of ash content in coal slime flotation tailings.

Abnormal Traffic Detection for Internet of Things Based on an Improved Residual Network

Along with the progression of Internet of Things(IoT)technology,network terminals are becoming continuously more intelligent.IoT has been widely applied in various scenarios,including urban infrastructure,transportation,industry,personal life,and other socio-economic fields.The introduction of deep learning has brought new security challenges,like an increment in abnormal traffic,which threatens network security.Insufficient feature extraction leads to less accurate classification results.In abnormal traffic detection,the data of network traffic is high-dimensional and complex.This data not only increases the computational burden of model training but also makes information extraction more difficult.To address these issues,this paper proposes an MD-MRD-ResNeXt model for abnormal network traffic detection.To fully utilize the multi-scale information in network traffic,a Multi-scale Dilated feature extraction(MD)block is introduced.This module can effectively understand and process information at various scales and uses dilated convolution technology to significantly broaden the model’s receptive field.The proposed Max-feature-map Residual with Dual-channel pooling(MRD)block integrates the maximum feature map with the residual block.This module ensures the model focuses on key information,thereby optimizing computational efficiency and reducing unnecessary information redundancy.Experimental results show that compared to the latest methods,the proposed abnormal traffic detection model improves accuracy by about 2%.

Underwater Image Enhancement Based on Multi-scale Adversarial Network

In this study,an underwater image enhancement method based on multi-scale adversarial network was proposed to solve the problem of detail blur and color distortion in underwater images.Firstly,the local features of each layer were enhanced into the global features by the proposed residual dense block,which ensured that the generated images retain more details.Secondly,a multi-scale structure was adopted to extract multi-scale semantic features of the original images.Finally,the features obtained from the dual channels were fused by an adaptive fusion module to further optimize the features.The discriminant network adopted the structure of the Markov discriminator.In addition,by constructing mean square error,structural similarity,and perceived color loss function,the generated image is consistent with the reference image in structure,color,and content.The experimental results showed that the enhanced underwater image deblurring effect of the proposed algorithm was good and the problem of underwater image color bias was effectively improved.In both subjective and objective evaluation indexes,the experimental results of the proposed algorithm are better than those of the comparison algorithm.

Residual Capacity of Friction⁃Type High⁃Strength Bolted T⁃stub Connection with Nut Corrosion Damage

Corrosion is a primary cause of the slippage of friction⁃type high⁃strength bolted(FHSB)T⁃stub connections.This paper attempts to quantify the residual capacity of FHSB T⁃stub connections with corroded nuts.Firstly,corrosion simulation tests were conducted on 48 manually cut nuts to find out the relationship between the damage degree of nut section and the residual clamping force(RCF)of bolt.Then,static load tests were carried out on 24 FHSB T⁃stub connections with nuts of different degrees of damage to obtain the failure modes.By finite⁃element(FE)models,a comparative analysis was performed on the initial friction load(IFL)and ultimate strength(US)of each connection with corroded nuts.Finally,the parameters of 96 FE models for FHSB T⁃stub connections were analyzed and used to derive the calculation formulas for the degree of damage for each nut and the IFL and US of each connection.The results show that the RCF decay of a bolt is a quadratic function of the equivalent radius loss ratio and the shear failure after nut corrosion;the IFL of each connection had a clear linear correlation with the RCF of the corresponding bolts,and the correlation depends on the applied load and static friction on connecting plate interface induced by the clamping force;the static friction had little impact on the US of the connection;the proposed IFL and US formulas can effectively derive the residual anti⁃slip capacity of FHSB T⁃stub connections from the degree of damage of the corroded nut section.The research results provide a scientific basis for the replacement and maintenance of corroded bolts of FHSB T⁃stub connections.

The Efficiency of Reduced Beam Section Connections for Reducing Residual Drifts in Moment Resisting Frames

In most framed structures anticipated deformations in accordance with current codes fall into acceptable limit states, whereas they go through substantial residual deformations in the aftermath of severe ground motions. These structures seem unsafe to occupants since static imminent instability in the immediate post-earthquake may be occurred. Moreover, rehabilitation costs of extensive residual deformations are not usually reasonable. Apparently, there is a lack of detailed knowledge related to reducing residual drift techniques when code-based seismic design is considered. In this paper, reduced beam section connections as a positive approach are taken action to mitigate the huge amount of residual drifts which are greatly amplified by P-Δ effects. To demonstrate the efficacy of RBS, a sixteen-story moment resisting frame is analyzed based on a suite of 8 single-component near field records which have been scaled according to the code provisions. The results are then processed to assess the effects of RBS detailing on drift profile, maximum drift, and residual drift. Besides, a special emphasis is given to estimate overall trend towards drift accumulation in each story in the presence of RBS assembly. A main conclusion is that using this connection predominantly alleviates the adverse effects of P-Δ on amplifying residual drifts.

Grasp Detection with Hierarchical Multi-Scale Feature Fusion and Inverted Shuffle Residual

Grasp detection plays a critical role for robot manipulation.Mainstream pixel-wise grasp detection networks with encoder-decoder structure receive much attention due to good accuracy and efficiency.However,they usually transmit the high-level feature in the encoder to the decoder,and low-level features are neglected.It is noted that low-level features contain abundant detail information,and how to fully exploit low-level features remains unsolved.Meanwhile,the channel information in high-level feature is also not well mined.Inevitably,the performance of grasp detection is degraded.To solve these problems,we propose a grasp detection network with hierarchical multi-scale feature fusion and inverted shuffle residual.Both low-level and high-level features in the encoder are firstly fused by the designed skip connections with attention module,and the fused information is then propagated to corresponding layers of the decoder for in-depth feature fusion.Such a hierarchical fusion guarantees the quality of grasp prediction.Furthermore,an inverted shuffle residual module is created,where the high-level feature from encoder is split in channel and the resultant split features are processed in their respective branches.By such differentiation processing,more high-dimensional channel information is kept,which enhances the representation ability of the network.Besides,an information enhancement module is added before the encoder to reinforce input information.The proposed method attains 98.9%and 97.8%in image-wise and object-wise accuracy on the Cornell grasping dataset,respectively,and the experimental results verify the effectiveness of the method.

Disease Recognition of Apple Leaf Using Lightweight Multi-Scale Network with ECANet

To solve the problem of difficulty in identifying apple diseases in the natural environment and the low application rate of deep learning recognition networks,a lightweight ResNet(LW-ResNet)model for apple disease recognition is proposed.Based on the deep residual network(ResNet18),the multi-scale feature extraction layer is constructed by group convolution to realize the compression model and improve the extraction ability of different sizes of lesion features.By improving the identity mapping structure to reduce information loss.By introducing the efficient channel attention module(ECANet)to suppress noise from a complex background.The experimental results show that the average precision,recall and F1-score of the LW-ResNet on the test set are 97.80%,97.92%and 97.85%,respectively.The parameter memory is 2.32 MB,which is 94%less than that of ResNet18.Compared with the classic lightweight networks SqueezeNet and MobileNetV2,LW-ResNet has obvious advantages in recognition performance,speed,parameter memory requirement and time complexity.The proposed model has the advantages of low computational cost,low storage cost,strong real-time performance,high identification accuracy,and strong practicability,which can meet the needs of real-time identification task of apple leaf disease on resource-constrained devices.

Multi-Scale Attention-Based Deep Neural Network for Brain Disease Diagnosis

Whole brain functional connectivity(FC)patterns obtained from resting-state functional magnetic resonance imaging(rs-fMRI)have been widely used in the diagnosis of brain disorders such as autism spectrum disorder(ASD).Recently,an increasing number of studies have focused on employing deep learning techniques to analyze FC patterns for brain disease classification.However,the high dimensionality of the FC features and the interpretation of deep learning results are issues that need to be addressed in the FC-based brain disease classification.In this paper,we proposed a multi-scale attention-based deep neural network(MSA-DNN)model to classify FC patterns for the ASD diagnosis.The model was implemented by adding a flexible multi-scale attention(MSA)module to the auto-encoder based backbone DNN,which can extract multi-scale features of the FC patterns and change the level of attention for different FCs by continuous learning.Our model will reinforce the weights of important FC features while suppress the unimportant FCs to ensure the sparsity of the model weights and enhance the model interpretability.We performed systematic experiments on the large multi-sites ASD dataset with both ten-fold and leaveone-site-out cross-validations.Results showed that our model outperformed classical methods in brain disease classification and revealed robust intersite prediction performance.We also localized important FC features and brain regions associated with ASD classification.Overall,our study further promotes the biomarker detection and computer-aided classification for ASD diagnosis,and the proposed MSA module is flexible and easy to implement in other classification networks.

Detection of influential nodes with multi-scale information

The identification of influential nodes in complex networks is one of the most exciting topics in network science.The latest work successfully compares each node using local connectivity and weak tie theory from a new perspective.We study the structural properties of networks in depth and extend this successful node evaluation from single-scale to multi-scale.In particular,one novel position parameter based on node transmission efficiency is proposed,which mainly depends on the shortest distances from target nodes to high-degree nodes.In this regard,the novel multi-scale information importance(MSII)method is proposed to better identify the crucial nodes by combining the network's local connectivity and global position information.In simulation comparisons,five state-of-the-art algorithms,i.e.the neighbor nodes degree algorithm(NND),betweenness centrality,closeness centrality,Katz centrality and the k-shell decomposition method,are selected to compare with our MSII.The results demonstrate that our method obtains superior performance in terms of robustness and spreading propagation for both real-world and artificial networks.

An Efficient Steganalysis Model Based on Multi-Scale LTP and Derivative Filters

Local binary pattern(LBP)is one of the most advanced image classification recognition operators and is commonly used in texture detection area.Research indicates that LBP also has a good application prospect in steganalysis.However,the existing LBP-based steganalysis algorithms are only capable to detect the least significant bit(LSB)and the least significant bit matching(LSBM)algorithms.To solve this problem,this paper proposes a steganalysis model called msdeLTP,which is based on multi-scale local ternary patterns(LTP)and derivative filters.The main characteristics of the msdeLTP are as follows:First,to reduce the interference of image content on features,the msdeLTP uses derivative filters to acquire residual images on which subsequent operations are based.Second,instead of LBP features,LTP features are extracted considering that the LTP feature can exhibit multiple variations in the relationship of adjacent pixels.Third,LTP features with multiple scales and modes are combined to show the relationship of neighbor pixels within different radius and along different directions.Analysis and simulation show that the msdeLTP uses only 2592-dimensional features and has similar detection accuracy as the spatial rich model(SRM)at the same time,showing the high steganalysis efficiency of the method.

Grinding/Cutting Technology and Equipment of Multi-scale Casting Parts

Multi-scale casting parts are important components of high-end equipment used in the aerospace,automobile manufacturing,shipbuilding,and other industries.Residual features such as parting lines and pouring risers that inevitably appear during the casting process are random in size,morphology,and distribution.The traditional manual processing method has disadvantages such as low efficiency,high labor intensity,and harsh working environment.Existing machine tool and serial robot grinding/cutting equipment do not easily achieve high-quality and high-efficiency removal of residual features due to poor dexterity and low stiffness,respectively.To address these problems,a five-degree-of-freedom(5-DoF)hybrid grinding/cutting robot with high dexterity and high stiffness is proposed.Based on it,three types of grinding/cutting equipment combined with offline programming,master-slave control,and other technologies are developed to remove the residual features of small,medium,and large casting parts.Finally,the advantages of teleoperation processing and other solutions are elaborated,and the difficulties and challenges are discussed.This paper reviews the grinding/cutting technology and equipment of casting parts and provides a reference for the research on the processing of multi-scale casting parts.

基于改进VGG16的自编码器视频异常检测算法

在使用自编码器结构的神经网络处理视频异常检测任务时,U-Net风格的自编码器由于编码器层数深度过浅,导致在面对复杂的数据集时,不能充分抽取更多有用的特征信息。同时,在训练模型时使用MSE(均方误差),仅考虑了预测帧与真实帧之间的像素级相似性,对于复杂场景,像素级相似性可能无法准确判断预测帧与真实帧之间的相似性。针对以上问题,对基于U-Net风格的自编码器进行改进,提出了一种使用改进的VGG16作为编码器的视频异常检测算法,同时在均方误差的基础上添加结构相似性(SSIM)损失函数。改进的VGG16去掉了全连接层,并加入了残差连接防止特征退化,添加SSIM在计算像素级相似性的同时计算图像的亮度、对比度和结构等方面的相似性来优化网络。实验结果表明,改进后的算法,在Ped2数据集上检测效果达到95.91%,在Avenue数据集上检测效果达到84.89%,与改进前的方法相比分别提高了0.80%和0.19%,验证了所提方法的有效性。

结合混合卷积和多尺度注意力的视频异常检测算法

基于U-net风格的无监督视频异常检测模型有着较好的检测效果,但由于普通卷积运算使用固有的局部特性,使U-Net风格的编码器无法有效地提取全局上下文信息,并且使用简单的跳跃连接无法获得有效的特征信息,使用的L2损失函数是仅考虑了像素级别的差异而无法捕捉图像的结构特征。对此提出了结合混合卷积和多尺度注意力的视频异常检测算法,并加入结构相似性损失函数(SSIM)优化模型。具体来说,在编码器最后一层添加混合卷积模块,混合空间和位置的特征来提取全局上下文信息。在编码器和解码器之间的跳跃连接中添加多尺度注意力模块,使模型能提取更有价值的特征,实现有效的跳跃连接。使用参数约束结构相似性损失函数与L2损失函数的权重,从而更准确地优化模型。实验结果表明,所提算法在UCSD-Ped2和CUHK Avenue公开数据集上的AUC指标达到96.7%和86.1%,与改进前的模型相比提高了1.6%和1.4%,证明了所提模型的有效性。

基于改进的空洞卷积UNet网络提取林地信息方法

为提升大范围林地信息提取的智能化程度及效率,提出了一种基于残差连接的空洞卷积UNet网络(RCD-UNet网络),将残差连接双卷积模块与传统UNet网络进行有机结合以提升模型性能,并以GF-2号高分辨率卫星遥感影像为数据源,开展了南京林地信息提取方法研究。结果表明:引入空洞金字塔池化层(ASPP)模块能够增强模型对上下文的感知能力,林地信息提取的总体精度为95.44%,Kappa系数为82.48%,满足高效、准确提取森林资源空间结构信息的需求,为森林资源管理与调查提供了技术支撑。

融合残差连接的图像语义分割方法

由于传统SegNet模型在采样过程中产生了大量信息损失,导致图像语义分割精度较低,为此提出了一种融合残差连接的新型编-解码器网络结构:文中引入了多残差连接策略,更为全面地保留了多尺度图像中包含的大量细节信息,降低还原降采样所带来的信息损失;为进一步加速网络训练的收敛效率,改善样本的不平衡问题,设计了一种带平衡因子的交叉熵损失函数,对正负样本不平衡现象予以针对性的优化,使得模型的训练更加高效;实验表明该方法较好地解决了语义分割中信息损失以及分割不准确的问题,与SegNet相比,本网络在Cityscapes数据集上进行精细标注的mIoU值提高了约13%。

基于改进AOD-Net的图像去雾算法

为了更好解决图像去雾后颜色失真、去雾不彻底和耗时等问题,提出了一种基于改进AOD-Net的图像去雾算法。首先,在原有的卷积模块中引入残差连接,并保留了第二个特征融合层第一层的特征信息,以增强网络的特征提取能力。其次,在第三个特征融合层后引入注意力模块,强化雾图中的关键特征信息,抑制无关背景干扰。最后,采用新的复合损失函数进行训练。实验结果表明,改进算法在公共数据集上的峰值信噪比提高了3.8 dB,结构相似性达到了93.6%。相较于其他去雾算法,该算法在去雾精度和处理效率方面均表现出色。

基于KCR-Informer的长期风电功率预测研究

准确的长期风电功率预测对电网系统稳定运行至关重要,传统预测方法在处理长序列预测时效果并不理想,近期研究表明Informer模型在长序列预测领域取得良好效果。然而,该模型在捕捉数据的局部特征以及处理网络层数堆叠问题上还有待改进。文章提出一种基于卡尔曼滤波器-卷积神经网络-残差网络-Informer(Kalman filter-convolutional neural network-residual network-informer,KCR-Informer)模型的长期风电功率预测方法,首先分析气象数据对风电功率的影响,使用卡尔曼滤波器对风电气象数据进行数据平滑处理,以减轻噪声对数据的影响,然后基于Informer模型建立风电功率预测模型,根据气象数据以及历史功率数据进行长期功率预测;在此基础上,引入卷积神经网络和残差连接模块,使模型能够更好的捕捉到局部特征,同时加快模型收敛,解决模型网络退化问题。算例的结果表明,与长短期记忆网络(long short-term memory,LSTM)算法、Transformer算法、Informer算法相比,文章方法在不同预测步长下的平均绝对误差(mean absolute error,MAE)降低5.7%~30%,均方误差(mean square error,MSE)降低8.3%~35%,长期风功率预测的精度得到提升,验证了模型的有效性。

一种基于多跳注意残差网络的调制识别算法

为了进一步提升通信信号调制识别的准确率,在ResNet网络的基础上提出一种基于多跳注意残差网络(Multi-skip Attention Residual Network,MARN)的调制识别方法。该方法利用提取不同特征的卷积核进行多跳连接构建3种残差块,进而构建多跳残差网络,提取信号的时域特征;加入CBAM(Convolutional Block Attention Module)注意力机制自适应地调整通道权重,加强信号特征的表征能力;采用自适配归一化(Switchable Normalization,SN)加速网络收敛;加入丢弃率为0.3的AlphaDropout层,提高算法的拟合能力,最终实现对通信信号端到端的分类识别。在RadioML2018.01a数据集上仿真实验,结果表明在信噪比为-10~15 dB下,MARN网络平均识别率达到63.3%,较ResNet网络的平均识别率提升3.7%。

基于密集残差连接的肺结节检测方法

针对目前基于深度学习的肺结节检测算法中不同深度与尺寸的特征信息间没有相互交流的问题,提出了一种基于密集残差连接的肺结节检测模型。本模型在3D U-Net网络的基础上引入密集连接,充分利用网络中肺结节特征图,实现不同层的特征信息的结合,提高结节特征的利用率;同时结合残差结构,避免了网络加深后出现的梯度消失问题;引入通道注意力机制,对不同通道的结节特征赋予权重,提高结节的识别率;在3D U-Net网络的编码解码部分间的跳跃连接中使用转置卷积,融合不同尺度与不同深度的特征。所提算法在肺结节公共数据集LUNA16上进行十折交叉验证,以无限制受试者操作特征为评价指标,实验结果表明,在假阳率为0.125、0.25、0.5、1、2、4、8这7个点上,平均敏感度为0.852,相较于基准模型提升5.5%。所提出的肺结节检测算法相比基准模型提高了检测敏感度,较好的实现对肺结节的检测。