基于改进Deformable-DETR的水下图像目标检测方法

针对由于水下复杂环境造成的目标检测效果较差、检测精度较低的问题,基于Deformable-DETR算法提出一种改进的水下目标检测算法Deformable-DETR-DA。使用空间注意力模块结合标准Transformer块设计了一个用于增加模型深度的深度特征金字塔(deep feature pyramid networks,DFPN)模块,将其嵌入到模型中提高模型对深层纹理信息的提取能力。使用注意力引导的方式对原模型中编码器部分进行改进,加强了对特征信息的聚合能力,提高了模型在复杂环境下的检测能力。针对URPC数据集,模型各交并比尺度的平均准确度(average precision,AP)为39.5%,相比原模型提升1%,与一些DETR(detection transformer)类的模型相比,不同目标尺度的平均准确度均有1%~4%左右的提高,表明改进的模型能够很好解决复杂环境的水下目标检测的问题。本文提出的模型可作为其他水下目标检测模型设计的参考。

基于改进Deformable DETR的无人机视频流车辆目标检测算法

针对无人机视频流检测中小目标数量多、因图像传输质量较低而导致的上下文语义信息不充分、传统算法融合特征推理速度慢、数据集类别样本不均衡导致的训练效果差等问题,提出一种基于改进Deformable DETR的无人机视频流车辆目标检测算法。在模型结构方面,该算法设计了跨尺度特征融合模块以增大感受野,提升小目标检测能力,并采用针对object_query的挤压-激励模块提升关键目标的响应值,减少重要目标的漏检与错检率;在数据处理方面,使用了在线困难样本挖掘技术,改善数据集中类别样本分布不均的问题。在UAVDT数据集上进行了实验,实验结果表明,改进后的算法相较于基线算法在平均检测精度上提升了1.5%,在小目标检测精度上提升了0.8%,并在保持参数量较少增长的情况下,维持了原有的检测速度。

A Self-Attention Based Dynamic Resource Management for Satellite-Terrestrial Networks

The satellite-terrestrial networks possess the ability to transcend geographical constraints inherent in traditional communication networks,enabling global coverage and offering users ubiquitous computing power support,which is an important development direction of future communications.In this paper,we take into account a multi-scenario network model under the coverage of low earth orbit(LEO)satellite,which can provide computing resources to users in faraway areas to improve task processing efficiency.However,LEO satellites experience limitations in computing and communication resources and the channels are time-varying and complex,which makes the extraction of state information a daunting task.Therefore,we explore the dynamic resource management issue pertaining to joint computing,communication resource allocation and power control for multi-access edge computing(MEC).In order to tackle this formidable issue,we undertake the task of transforming the issue into a Markov decision process(MDP)problem and propose the self-attention based dynamic resource management(SABDRM)algorithm,which effectively extracts state information features to enhance the training process.Simulation results show that the proposed algorithm is capable of effectively reducing the long-term average delay and energy consumption of the tasks.

Missing Value Imputation for Radar-Derived Time-Series Tracks of Aerial Targets Based on Improved Self-Attention-Based Network

The frequent missing values in radar-derived time-series tracks of aerial targets(RTT-AT)lead to significant challenges in subsequent data-driven tasks.However,the majority of imputation research focuses on random missing(RM)that differs significantly from common missing patterns of RTT-AT.The method for solving the RM may experience performance degradation or failure when applied to RTT-AT imputation.Conventional autoregressive deep learning methods are prone to error accumulation and long-term dependency loss.In this paper,a non-autoregressive imputation model that addresses the issue of missing value imputation for two common missing patterns in RTT-AT is proposed.Our model consists of two probabilistic sparse diagonal masking self-attention(PSDMSA)units and a weight fusion unit.It learns missing values by combining the representations outputted by the two units,aiming to minimize the difference between the missing values and their actual values.The PSDMSA units effectively capture temporal dependencies and attribute correlations between time steps,improving imputation quality.The weight fusion unit automatically updates the weights of the output representations from the two units to obtain a more accurate final representation.The experimental results indicate that,despite varying missing rates in the two missing patterns,our model consistently outperforms other methods in imputation performance and exhibits a low frequency of deviations in estimates for specific missing entries.Compared to the state-of-the-art autoregressive deep learning imputation model Bidirectional Recurrent Imputation for Time Series(BRITS),our proposed model reduces mean absolute error(MAE)by 31%~50%.Additionally,the model attains a training speed that is 4 to 8 times faster when compared to both BRITS and a standard Transformer model when trained on the same dataset.Finally,the findings from the ablation experiments demonstrate that the PSDMSA,the weight fusion unit,cascade network design,and imputation loss enhance imputation performance and confirm the efficacy of our design.

Aerial target threat assessment based on gated recurrent unit and self-attention mechanism

Aerial threat assessment is a crucial link in modern air combat, whose result counts a great deal for commanders to make decisions. With the consideration that the existing threat assessment methods have difficulties in dealing with high dimensional time series target data, a threat assessment method based on self-attention mechanism and gated recurrent unit(SAGRU) is proposed. Firstly, a threat feature system including air combat situations and capability features is established. Moreover, a data augmentation process based on fractional Fourier transform(FRFT) is applied to extract more valuable information from time series situation features. Furthermore, aiming to capture key characteristics of battlefield evolution, a bidirectional GRU and SA mechanisms are designed for enhanced features.Subsequently, after the concatenation of the processed air combat situation and capability features, the target threat level will be predicted by fully connected neural layers and the softmax classifier. Finally, in order to validate this model, an air combat dataset generated by a combat simulation system is introduced for model training and testing. The comparison experiments show the proposed model has structural rationality and can perform threat assessment faster and more accurately than the other existing models based on deep learning.

Modeling Near-Field Impulsive Waves Generated by Deformable Landslide Using the HBP Model Based on the SPH Method

Landslide-generated impulsive waves(LGWs)in reservoir areas can seriously threaten waterway safety as well as hu-man life and properties around the two side slopes.The risk reduction and mitigation of such a hazard require the accurate prediction of near-field wave characteristics,such as wave amplitude and run-up.However,near-field LGW involves complicated fluid-solid interactions.Furthermore,the wave characteristics are closely related to various aspects,including the geometry and physical features of the slide,river,and body of water.However,the empirical or analytical methods used for rough estimation cannot derive accurate results,especially for deformable landslides,due to their significant geometry changes during the sliding process.In this study,the near-field waves generated by deformable landslides were simulated by smoothed particle hydrodynamics(SPH)based on multi-phase flow.The deformable landslides were generalized as a kind of viscous flow by adopting the Herschel-Bulkley-Papanastasiou(HBP)-based nonNewtonian rheology model.The HBP model is capable of producing deformable landslide dynamics even though the high-speed sliding process is involved.In this study,an idealized experiment case originating from Lituya LGW and a practical case of Gongjiafang LGW were reproduced for verification and demonstration.The simulation results of both cases show satisfactory consistency with the experiment/investigation data in terms of landslide movement and near-field impulsive wave characteristics,thus indicating the applicability and accuracy of the proposed method.Finally,the effects of the HBP model’s rheological parameters on the landslide dynamics and near-field wave characteristics are discussed,providing a parameter calibration method along with sug-gestions for further applications.

Deformable Catalytic Material Derived from Mechanical Flexibility for Hydrogen Evolution Reaction

Deformable catalytic material with excellent flexible structure is a new type of catalyst that has been applied in various chemical reactions,especially electrocatalytic hydrogen evolution reaction(HER).In recent years,deformable catalysts for HER have made great progress and would become a research hotspot.The catalytic activities of deformable catalysts could be adjustable by the strain engineering and surface reconfiguration.The surface curvature of flexible catalytic materials is closely related to the electrocatalytic HER properties.Here,firstly,we systematically summarized self-adaptive catalytic performance of deformable catalysts and various micro–nanostructures evolution in catalytic HER process.Secondly,a series of strategies to design highly active catalysts based on the mechanical flexibility of lowdimensional nanomaterials were summarized.Last but not least,we presented the challenges and prospects of the study of flexible and deformable micro–nanostructures of electrocatalysts,which would further deepen the understanding of catalytic mechanisms of deformable HER catalyst.

基于Deformable DETR的自然场景任意形状文本检测

自然场景下的文本区域形状复杂多变,直接使用轮廓坐标描述文本区域会使得建模不充分,导致文本检测准确性低。针对自然场景下文本区域不规则的问题,提出了一种基于Deformable DETR的任意形状文本检测模型,不同于传统的直接预测轮廓点的方法,使用B-样条对文字区域进行建模使得文本轮廓平滑精确的同时减少了需要预测的参数。提出的文本检测模型无需手工设计锚点、区域建议等组件,极大地简化了模型设计并提高了通用性。提出的模型在无需额外数据集的情况下在任意形状文本数据集CTW1500和Total-Text上的平均精度(F值)分别达到了85.4%和85.0%,证明了模型的有效性。

基于Deformable DETR的红外图像目标检测方法研究

基于Transformer架构的DETR系列网络在计算机视觉目标检测领域不断刷新目标检测的精度与速度。然而,基于红外图像的非合作目标检测的应用环境复杂,而且红外图像质量较差。针对该问题,提出了一种新的以Deformable DETR算法为基线的具有高检测精度的目标检测算法:首先设计了对红外图像进行图像增强处理的图像增强模块CLAHE-GB,并将其与Deformable DETR进行有机结合;然后在大型通用数据集上进行预训练;最后引入数据增强和迁移学习方法在自制的空中飞行物小型红外图像数据集中对检测头网络参数进行再训练,并对结果进行分析。结果表明:所提出的算法对红外图像数据具有较好的图像增强效果和检测精度。

High-resolution visible imaging with piezoelectric deformable secondary mirror: experimental results at the 1.8-m adaptive telescope

Integrating deformable mirrors within the optical train of an adaptive telescope was one of the major innovations in astronomical observation technology,distinguished by its high optical throughput,reduced optical surfaces,and the incorporation of the deformable mirror.Typically,voice-coil actuators are used,which require additional position sensors,internal control electronics,and cooling systems,leading to a very complex structure.Piezoelectric deformable secondary mirror technologies were proposed to overcome these problems.Recently,a high-order piezoelectric deformable secondary mirror has been developed and installed on the 1.8-m telescope at Lijiang Observatory in China to make it an adaptive telescope.The system consists of a 241-actuator piezoelectric deformable secondary mirror,a 192-sub-aperture Shack-Hartmann wavefront sensor,and a multi-core-based real-time controller.The actuator spacing of the PDSM measures 19.3 mm,equivalent to approximately 12.6 cm when mapped onto the primary mirror,significantly less than the voicecoil-based adaptive telescopes such as LBT,Magellan and VLT.As a result,stellar images with Strehl ratios above 0.49 in the R band have been obtained.To our knowledge,these are the highest R band images captured by an adaptive telescope with deformable secondary mirrors.Here,we report the system description and on-sky performance of this adaptive telescope.

CFSA-Net:Efficient Large-Scale Point Cloud Semantic Segmentation Based on Cross-Fusion Self-Attention

Traditional models for semantic segmentation in point clouds primarily focus on smaller scales.However,in real-world applications,point clouds often exhibit larger scales,leading to heavy computational and memory requirements.The key to handling large-scale point clouds lies in leveraging random sampling,which offers higher computational efficiency and lower memory consumption compared to other sampling methods.Nevertheless,the use of random sampling can potentially result in the loss of crucial points during the encoding stage.To address these issues,this paper proposes cross-fusion self-attention network(CFSA-Net),a lightweight and efficient network architecture specifically designed for directly processing large-scale point clouds.At the core of this network is the incorporation of random sampling alongside a local feature extraction module based on cross-fusion self-attention(CFSA).This module effectively integrates long-range contextual dependencies between points by employing hierarchical position encoding(HPC).Furthermore,it enhances the interaction between each point’s coordinates and feature information through cross-fusion self-attention pooling,enabling the acquisition of more comprehensive geometric information.Finally,a residual optimization(RO)structure is introduced to extend the receptive field of individual points by stacking hierarchical position encoding and cross-fusion self-attention pooling,thereby reducing the impact of information loss caused by random sampling.Experimental results on the Stanford Large-Scale 3D Indoor Spaces(S3DIS),Semantic3D,and SemanticKITTI datasets demonstrate the superiority of this algorithm over advanced approaches such as RandLA-Net and KPConv.These findings underscore the excellent performance of CFSA-Net in large-scale 3D semantic segmentation.

Clothing Parsing Based on Multi-Scale Fusion and Improved Self-Attention Mechanism

Due to the lack of long-range association and spatial location information,fine details and accurate boundaries of complex clothing images cannot always be obtained by using the existing deep learning-based methods.This paper presents a convolutional structure with multi-scale fusion to optimize the step of clothing feature extraction and a self-attention module to capture long-range association information.The structure enables the self-attention mechanism to directly participate in the process of information exchange through the down-scaling projection operation of the multi-scale framework.In addition,the improved self-attention module introduces the extraction of 2-dimensional relative position information to make up for its lack of ability to extract spatial position features from clothing images.The experimental results based on the colorful fashion parsing dataset(CFPD)show that the proposed network structure achieves 53.68%mean intersection over union(mIoU)and has better performance on the clothing parsing task.

Liquefaction susceptibility and deformation characteristics of saturated coral sandy soils subjected to cyclic loadings-a critical review

Coral sandy soils widely exist in coral island reefs and seashores in tropical and subtropical regions.Due to the unique marine depositional environment of coral sandy soils,the engineering characteristics and responses of these soils subjected to monotonic and cyclic loadings have been a subject of intense interest among the geotechnical and earthquake engineering communities.This paper critically reviews the progress of experimental investigations on the undrained behavior of coral sandy soils under monotonic and cyclic loadings over the last three decades.The focus of coverage includes the contractive-dilative behavior,the pattern of excess pore-water pressure(EPWP)generation and the liquefaction mechanism and liquefaction resistance,the small-strain shear modulus and strain-dependent shear modulus and damping,the cyclic softening feature,and the anisotropic characteristics of undrained responses of saturated coral sandy soils.In particular,the advances made in the past decades are reviewed from the following aspects:(1)the characterization of factors that impact the mechanism and patterns of EPWP build-up;(2)the identification of liquefaction triggering in terms of the apparent viscosity and the average flow coefficient;(3)the establishment of the invariable form of strain-based,stress-based,or energy-based EPWP ratio formulas and the unique relationship between the new proxy of liquefaction resistance and the number of cycles required to reach liquefaction;(4)the establishment of the invariable form of the predictive formulas of small strain modulus and strain-dependent shear modulus;and(5)the investigation on the effects of stress-induced anisotropy on liquefaction susceptibility and dynamic deformation characteristics.Insights gained through the critical review of these advances in the past decades offer a perspective for future research to further resolve the fundamental issues concerning the liquefaction mechanism and responses of coral sandy sites subjected to cyclic loadings associated with seismic events in marine environments.

Design of the yielding support used highly deformable elements for a tunnel excavated in squeezing rock

Yielding support is often used in the squeezing tunnel to prevent damage to the lining induced by large deformation of the surrounding rock.Highly Deformable Elements(HDE)which is often installed along the circumferential direction of the shotcrete lining is a common type of yielding support.To determine the yield parameters of HDE,the support characteristic of the lining using HDE and the ground pressure considering strain-softening of soft rock were analyzed by an analytical method.The analytical solution showed that when considering the strain-softening of squeezing ground,the ground pressure has a non-zero minimum value.The minimum value of ground stress can be used to determine the constant yield stress of the HDE,and the corresponding deformation of the minimum ground pressure can be used to determine the deformation capacity of the HDE.Based on the variation in the design constant yield stress and yield displacement of HDE with the in-situ stress and the mechanical parameters of the soft rock,equations were proposed for determining of the yield parameters of the HDE.

Deformation monitoring of long-span railway bridges based on SBAS-InSAR technology

The deformation monitoring of long-span railway bridges is significant to ensure the safety of human life and property.The interferometric synthetic aperture radar(In SAR)technology has the advantage of high accuracy in bridge deformation monitoring.This study monitored the deformation of the Ganjiang Super Bridge based on the small baseline subsets(SBAS)In SAR technology and Sentinel-1A data.We analyzed the deformation results combined with bridge structure,temperature,and riverbed sediment scouring.The results are as follows:(1)The Ganjiang Super Bridge area is stable overall,with deformation rates ranging from-15.6 mm/yr to 10.7 mm/yr(2)The settlement of the Ganjiang Super Bridge deck gradually increases from the bridge tower toward the main span,which conforms to the typical deformation pattern of a cable-stayed bridge.(3)The sediment scouring from the riverbed cause the serious settlement on the bridge’s east side compared with that on the west side.(4)The bridge deformation negatively correlates with temperature,with a faster settlement at a higher temperature and a slow rebound trend at a lower temperature.The study findings can provide scientific data support for the health monitoring of long-span railway bridges.

基于Self-Attention的方面级情感分析方法研究

针对传统模型在细粒度的方面级情感分析上的不足,如RNN会遇到长距离依赖的问题,且模型不能并行计算;CNN的输出通常包含池化层,特征向量经过池化层的运算后会丢失相对位置信息和一些重要特征,且CNN没有考虑到文本的上下文信息。本文提出了一种Light-Transformer-ALSC模型,基于Self-Attention机制,且运用了交互注意力的思想,对方面词和上下文使用不同的注意力模块提取特征,细粒度地对文本进行情感分析,在SemEval2014 Task 4数据集上的实验结果表明本文模型的效果优于大部分仅基于LSTM的模型。除基于BERT的模型外,在Laptop数据集上准确率提高了1.3%~5.3%、在Restaurant数据集上准确率提高了2.5%~5.5%;对比基于BERT的模型,在准确率接近的情况下模型参数量大大减少。

Effect of deformation parameters on the austenite dynamic recrystallization behavior of a eutectoid pearlite rail steel

Understandings of the effect of hot deformation parameters close to the practical production line on grain refinement are crucial for enhancing both the strength and toughness of future rail steels.In this work,the austenite dynamic recrystallization(DRX)behaviors of a eutectoid pearlite rail steel were studied using a thermo-mechanical simulator with hot deformation parameters frequently employed in rail production lines.The single-pass hot deformation results reveal that the prior austenite grain sizes(PAGSs)for samples with different deformation reductions decrease initially with an increase in deformation temperature.However,once the deformation temperature is beyond a certain threshold,the PAGSs start to increase.It can be attributed to the rise in DRX volume fraction and the increase of DRX grain with deformation temperature,respectively.Three-pass hot deformation results show that the accumulated strain generated in the first and second deformation passes can increase the extent of DRX.In the case of complete DRX,PAGS is predominantly determined by the deformation temperature of the final pass.It suggests a strategic approach during industrial production where part of the deformation reduction in low temperature range can be shifted to the medium temperature range to release rolling mill loads.

Mechanical and magnetocaloric adjustable properties of Fe3O4/PET deformed nanoparticle film

The flexibility of nanoparticle films is a topic of rapidly growing interest in both scientific and engineering researches due to their numerous potential applications in a broad range of wearable electronics and biomedical devices.This article presents the elucidation of the properties of nanoparticle films.Here,a flexible film is fabricated based on polyethylene terephthalate(PET)and magnetic iron oxide at the nanoscale using layer-by-layer technology.The 2D thin flexible film material can be bent at different angles from 0°to 360°.With an increase in elastic deformation angles,the magnetocaloric effect of the film gradually increases in the alternating magnetic field.The test results from a vibrating sample magnetometer and a low-frequency impedance analyzer demonstrate that the film has a good magnetic response and anisotropy.The magnetocaloric effect and magnetic induction effect are controlled by deformation,providing a new idea for the application of elastic films.It combines the flexibility of the nanoparticle PET substrate and,in the future,it may be used for skin adhesion for administration and magnetic stimulation control.

Mechanical Behavior and Microstructure Evolution during Tensile Deformation of Twinning Induced Plasticity Steel Processed by Warm Forgings

The mechanical behavior and microstructural evolution of an Fe-30Mn-3Al-3Si twinninginduced plasticity(TWIP)steel processed using warm forging was investigated.It is found that steel processed via warm forging improves comprehensive mechanical properties compared to the TWIP steel processed via cold rolling,with a high tensile strength(R_(m))of 793 MPa,a yield strength(R_(P))of 682 MPa,an extremely large R_(P)/R_(m)ratio as high as 0.86 as well as an excellent elongation rate of 46.8%.The microstructure observation demonstrates that steel processed by warm forging consists of large and elongated grains together with fine,equiaxed grains.Complicated micro-defect configurations were also observed within the steel,including dense dislocation networks and a few coarse deformation twins.As the plastic deformation proceeds,the densities of dislocations and deformation twins significantly increase.Moreover,a great number of slip lines could be observed in the elongated grains.These findings reveal that a much more dramatic interaction between microstructural defect and dislocations glide takes place in the forging sample,wherein the fine and equiaxed grains propagated dislocations more rapidly,together with initial defect configurations,are responsible for enhanced strength properties.Meanwhile,larger,elongated grains with more prevalently activated deformation twins result in high plasticity.

Unraveling engineering disturbance effects on deformation in red-bed mudstone railway cuttings:incorporating crack-facilitated moisture diffusion

Red-bed mudstone, prevalent in southwest China, poses a formidable challenge due to its hydrophilic clay minerals, resulting in expansion, deformation, and cracking upon exposure to moisture. This study addresses uplift deformation disasters in high-speed railways by employing a moisture diffusion-deformation-fracture coupling model based on the finite-discrete element method(FDEM). The model integrates the influence of cracks on moisture diffusion. The investigation into various excavation depths reveals a direct correlation between depth and the formation of tensile cracks at the bottom of the railway cutting. These cracks expedite moisture migration, significantly impacting the temporal and spatial evolution of the moisture field. Additionally, crack expansion dominates hygroscopic deformation, with the lateral coordinate of the crack zone determining peak vertical displacement. Furthermore, key factors influencing deformation in railway cuttings, including the swelling factor and initial moisture content at the bottom of the cutting, are explored. The number of tensile and shear cracks increases with greater excavation depth, particularly concerning shear cracks. Higher swelling factors and initial moisture contents result in an increased total number of cracks, predominantly shear cracks. Numerical calculations provide valuable insights, offering a scientific foundation and directional guidance for the precise prevention, control, prediction, and comprehensive treatment of mudstone-related issues in high-speed railways.