Source Camera Identification Algorithm Based on Multi-Scale Feature Fusion

The widespread availability of digital multimedia data has led to a new challenge in digital forensics.Traditional source camera identification algorithms usually rely on various traces in the capturing process.However,these traces have become increasingly difficult to extract due to wide availability of various image processing algorithms.Convolutional Neural Networks(CNN)-based algorithms have demonstrated good discriminative capabilities for different brands and even different models of camera devices.However,their performances is not ideal in case of distinguishing between individual devices of the same model,because cameras of the same model typically use the same optical lens,image sensor,and image processing algorithms,that result in minimal overall differences.In this paper,we propose a camera forensics algorithm based on multi-scale feature fusion to address these issues.The proposed algorithm extracts different local features from feature maps of different scales and then fuses them to obtain a comprehensive feature representation.This representation is then fed into a subsequent camera fingerprint classification network.Building upon the Swin-T network,we utilize Transformer Blocks and Graph Convolutional Network(GCN)modules to fuse multi-scale features from different stages of the backbone network.Furthermore,we conduct experiments on established datasets to demonstrate the feasibility and effectiveness of the proposed approach.

Neutronics analysis of a subcritical blanket system driven by a gas dynamic trap-based fusion neutron source for ^(99)Mo production

Gamma-emitting radionuclide ^(99m)Tc is globally used for the diagnosis of various pathological conditions owing to its ideal single-photon emission computed tomography (SPECT) characteristics.However,the short half-life of ^(99m)Tc (T_(1/2)=6 h)makes it difficult to store or transport.Thus,the production of ^(99m)Tc is tied to its parent radionuclide ^(99)Mo (T_(1/2)=66 h).The major production paths are based on accelerators and research reactors.The reactor process presents the potential for nuclear proliferation owing to its use of highly enriched uranium (HEU).Accelerator-based methods tend to use deuterium–tritium(D–T) neutron sources but are hindered by the high cost of tritium and its challenging operation.In this study,a new ^(99)Mo production design was developed based on a deuterium–deuterium (D–D) gas dynamic trap fusion neutron source (GDT-FNS) and a subcritical blanket system (SBS) assembly with a low-enriched uranium (LEU) solution.GDT-FNS can provide a relatively high-neutron intensity,which is one of the advantages of ^(99)Mo production.We provide a Monte Carlo-based neutronics analysis covering the calculation of the subcritical multiplication factor (k_(s)) of the SBS,optimization design for the reflector,shielding layer,and ^(99)Mo production capacity.Other calculations,including the neutron flux and nuclear heating distributions,are also provided for an overall evaluation of the production system.The results demonstrated that the SBS meets the nuclear critical safety design requirement (k_(s)<0.97) and maintained a high ^(99)Mo production capacity.The proposed system can generate approximately 157 Ci ^(99)Mo for a stable 24 h operation with a neutron intensity of 1×10^(14) n/s,which can meet 50%of China’s demand in 2025.

Effect of Fusion Neutron Source Numerical Models on Neutron Wall Loading in a D-D Tokamak Device

Effect of various spatial and energy distributions of fusion neutron sourceon the calculation of neutron wall loading of Tokamak D-D fusion device has been investigated bymeans of the 3-D Monte Carlo code MCNP. A realistic Monte Carlo source model was developed based onthe accurate representation of the spatial distribution and energy spectrum of fusion neutrons tosolve the complicated problem of tokamak fusion neutron source modelling. The results show thatthose simplified source models will introduce significant uncertainties. For accurate estimation ofthe key nuclear responses of the tokamak design and analyses, the use of the realistic source isrecommended. In addition, the accumulation of tritium produced during D-D plasma operation should becarefully considered.

A Fusion Neutron Source Driven Sub-Critical Nuclear Energy System: A Way for Early Application of Fusion Technology

This paper proposes a sub-critical nuclear energy system driven by fusion neutron source, FDS, which can be used to transmute long-lived radioactive wastes and to produce fissile nuclear fuel as a way for early application of fusion technology. The necessity and feasibility to develop that system in China are illustrated on the basis of prediction of the demand of energy source in the first half of the 21th century, the status of current fission energy supply and the progress in fusion technology in the world. The characteristics of fusion neutron driver and the potential for transmutation of long-lived nuclear wastes and breeding of fissile nuclear fuel in a blanket are analyzed. A scenario of development steps is proposed.

Laser ion source for heavy ion inertial fusion

The proposed heavy ion inertial fusion(HIF)scenarios require ampere class low charge state ion beams of heavy species.The laser ion source(LIS)is recognized as one of the promising candidates of ion beam providers,since it can deliver high brightness heavy ion beams to accelerators.The design of LIS for the HIF depends on the accelerator structure and accelerator complex following the source.In this article,we discuss the specifications and design of an appropriate LIS assuming two major types of the accelerators:radio frequency(RF)high quality factor cavity type and non-resonant induction core type.We believe that a properly designed LIS would satisfy the requirements of both types,while some issues need to be verified experimentally.

Physics Analysis and Optimization Studies for a Fusion Neutron Source Based on a Gas Dynamic Trap

To further investigate the fusion neutron source based on a gas dynamic trap （GDT）, characteristics of the GDT were analyzed and physics analyses were made for a fusion neutron source based on the GDT concept. The prior design of a GDT-based fusion neutron source was optimized based on a refreshed understanding of GDT operation. A two-step progressive development route of a GDT-based fusion neutron source was suggested. Potential applications of GDT are discussed. Preliminary analyses show that a fusion neutron source based on the GDT concept is suitable for plasma-material interaction research, fusion material and subcomponent testing, and capable of driving a proof-of-principle fusion fission hybrid experimental facility.

A multi-source data fusion modeling method for debris flow prevention engineering

The Digital Elevation Model(DEM)data of debris flow prevention engineering are the boundary of a debris flow prevention simulation,which provides accurate and reliable DEM data and is a key consideration in debris flow prevention simulations.Thus,this paper proposes a multi-source data fusion method.First,we constructed 3D models of debris flow prevention using virtual reality technology according to the relevant specifications.The 3D spatial data generated by 3D modeling were converted into DEM data for debris flow prevention engineering.Then,the accuracy and applicability of the DEM data were verified by the error analysis testing and fusion testing of the debris flow prevention simulation.Finally,we propose the Levels of Detail algorithm based on the quadtree structure to realize the visualization of a large-scale disaster prevention scene.The test results reveal that the data fusion method controlled the error rate of the DEM data of the debris flow prevention engineering within an allowable range and generated 3D volume data(obj format)to compensate for the deficiency of the DEM data whereby the 3D internal entity space is not expressed.Additionally,the levels of detailed method can dispatch the data of a large-scale debris flow hazard scene in real time to ensure a realistic 3D visualization.In summary,the proposed methods can be applied to the planning of debris flow prevention engineering and to the simulation of the debris flow prevention process.

3D fluid model analysis on the generation of negative hydrogen ions for negative ion source of NBI

A radio-frequency(RF) inductively coupled negative hydrogen ion source(NHIS) has been adopted in the China Fusion Engineering Test Reactor(CFETR) to generate negative hydrogen ions.By incorporating the level-lumping method into a three-dimensional fluid model,the volume production and transportation of H^(-) in the NHIS,which consists of a cylindrical driver region and a rectangular expansion chamber,are investigated self-consistently at a large input power(40 k W) and different pressures(0.3–2.0 Pa).The results indicate that with the increase of pressure,the H^(-) density at the bottom of the expansion region first increases and then decreases.In addition,the effect of the magnetic filter is examined.It is noteworthy that a significant increase in the H^(-) density is observed when the magnetic filter is introduced.As the permanent magnets move towards the driver region,the H^(-) density decreases monotonically and the asymmetry is enhanced.This study contributes to the understanding of H-distribution under various conditions and facilitates the optimization of volume production of negative hydrogen ions in the NHIS.

Locality preserving fusion of multi-source images for sea-ice classification

We present a novel sea-ice classification framework based on locality preserving fusion of multi-source images information.The locality preserving fusion arises from two-fold,i.e.,the local characterization in both spatial and feature domains.We commence by simultaneously learning a projection matrix,which preserves spatial localities,and a similarity matrix,which encodes feature similarities.We map the pixels of multi-source images by the projection matrix to a set fusion vectors that preserve spatial localities of the image.On the other hand,by applying the Laplacian eigen-decomposition to the similarity matrix,we obtain another set of fusion vectors that preserve the feature local similarities.We concatenate the fusion vectors for both spatial and feature locality preservation and obtain the fusion image.Finally,we classify the fusion image pixels by a novel sliding ensemble strategy,which enhances the locality preservation in classification.Our locality preserving fusion framework is effective in classifying multi-source sea-ice images(e.g.,multi-spectral and synthetic aperture radar(SAR)images)because it not only comprehensively captures the spatial neighboring relationships but also intrinsically characterizes the feature associations between different types of sea-ices.Experimental evaluations validate the effectiveness of our framework.

Image Processing on Geological Data in Vector Format and Multi-Source Spatial Data Fusion

The geological data are constructed in vector format in geographical information system (GIS) while other data such as remote sensing images, geographical data and geochemical data are saved in raster ones. This paper converts the vector data into 8 bit images according to their importance to mineralization each by programming. We can communicate the geological meaning with the raster images by this method. The paper also fuses geographical data and geochemical data with the programmed strata data. The result shows that image fusion can express different intensities effectively and visualize the structure characters in 2 dimensions. Furthermore, it also can produce optimized information from multi-source data and express them more directly.

Artificial Intelligence Integration with Energy Sources (Renewable and Non-renewable)

In the past decade or so,AI(artificial intelligence)technology has been growing with such a mesmerizing speed that today its presence in almost any industry that deals with any huge sheer volume of data is taking advantage of AI by integrating it into their day-to-day operation.Meanwhile,seven billion people worldwide shape the world’s energy system and directly impact the fundamental drivers of energy,both renewable and non-renewable sources,to meet the demand for electricity from them.These energy sources can be reached from nature such as solar,wind,etc.,and human-made such as NPPs(nuclear power plants)in the form of either fission as an old technology since the Manhattan project and in the near future as fusion in the form of magnetic or inertial confinements.Meanwhile,AI controlling nuclear reactors are about to happen.The basic idea is to apply AI with its two subset components as ML(machine learning),and DL(deep learning)techniques to go through the mountains of data that come from a reactor,spot patterns in it,and calling them to the unit’s human attention operators is not invadable either.Designers of such nuclear reactors will combine simulation and real-world data,comparing scenarios from each to develop“confidence[in]what they can predict and what is the range of uncertainty of their prediction”.Adding that,in the end,the operator will make the final decisions in order to keep these power plants safe while they are in operation and how to secure them against cyber-attack natural or human-made disasters.In this short communication article,we would like to see how we can prove some of these concepts;then a NPP manufacturer can pick it up and use it in their designs of a new generation of these reactors.

The Combination Operator of Information Sources by a New Expressive Matrix

The multi-sensors fusion refers to the synergistic combination of sensory data from multiple sensors to provide more accurate and reliable information. The potential benefits of the Fusion are multi-sensors’ redundancy and extra information acquired. The fusion of redundant information can reduce the overall uncertainty and thus helps to provide information specified more precisely. Several sensors providing redundant information can also be used to increase reliability in the case of error, omission or failure of sensors. The combination operators are exponential and are more complex in terms of calculation;the Dempster-Shafer operator is exponential for more than three (3) information sources?[1] [2]. Our work focuses on the definition of another formulation of this operation, and puts it in a matrix form to illuminate the computational complexity, more precision guaranty and a minimal execution time. We propose to use each information source in a form of a matrix, which contains 0 value in lines that do not contain the masses (m(Ai) = 0) or once m(Ai) is not null (m(Ai) ≠ 0). The use of this expressed matrix attempts to ameliorate Dempster-Shafer operator via initialing either a criterion or criteria sources’ solution, increasing the efficiency of the Dempster-Shafer operator and facilitates the combination among the sources. We evaluate our approach by conducting a case study for showing the effectiveness of this matrix.

The Parameters of the Stellarator as a Neutron Source for a Subcritical Reactor

The possibility of developing a stellarator-based neutron source designed for the nuclear reaction initiation in the blanket of hybrid reactor is studied. An analog of the Large Helical Device （LHD） stellarator design, with linear dimensions increased by a factor of 1.5 is taken for the magnetic system. Plasma parameters and the deuterium-tritium （DT） mixture fusion power are calculated using the space-time numerical code under the assumption of the neoclassical transport in the ambipolarity regime. Using the 10 MW plasma heating sources, it is possible to obtain the DT fusion power of one-to-two tens MW.

Explanation of Cold Nuclear Fusion and Biotransmutations

Low energy nuclear reactions are possible in condensed matter because of image forces. They result from induced charges at the surface of metals or very polarizable media. The height and width of the Coulomb barrier in free space can thus be reduced. Nuclear fusion requires also the formation of a compound nucleus in one of its excited states, but two deuterons yield an α particle that has 2 excited states. They are respectively accessible at high or low energies. Since the reduction of the Coulomb barrier depends on the local curvature of the interface, cold fusion becomes autocatalytic, but heat production is controllable. Even microbes, plants and animals can produce transmutations. They are also due to image forces. This solves a basic problem in nuclear physics and there are possible applications: facilitated synthesis of superheavy elements and development of a new type of energy sources. They are moderate, but safe.

基于Transformer和动态3D卷积的多源遥感图像分类

多源遥感数据具有互补性和协同性,近年来,基于深度学习的方法已经在多源遥感图像分类中取得了一定进展,但当前方法仍面临关键难题,如多源遥感图像特征表达不一致,融合困难,基于静态推理范式的神经网络缺乏对不同类别地物的适应性。为解决上述问题,提出了基于跨模态Transformer和多尺度动态3D卷积的多源遥感图像分类模型。为提高多源特征表达的一致性,设计了基于Transformer的融合模块,借助其强大的注意力建模能力挖掘高光谱和LiDAR数据特征之间的相互作用;为提高特征提取方法对不同地物类别的适应性,设计了多尺度动态3D卷积模块,将输入特征的多尺度信息融入卷积核的调制,提高卷积操作对不同地物的适应性。采用多源遥感数据集Houston和Trento对所提方法进行验证,实验结果表明:所提方法在Houston和Trento数据集上总体准确率分别达到94.60%和98.21%,相比MGA-MFN等主流方法,总体准确率分别至少提升0.97%和0.25%,验证了所提方法可有效提升多源遥感图像分类的准确率。

自研28 GHz/50 kW回旋管实现长时间连续运行

介绍了中国工程物理研究院应用电子学研究所针对磁约束聚变装置电子回旋共振加热(ECRH)系统、重离子加速器电子回旋共振(ECR)离子源以及前沿科技探索应用研制的28 GHz/50 kW连续波回旋管最新实验结果。研究团队在2019年该回旋管实现50 kW/30 s运行的基础上,通过结构优化和稳定性设计验证,最终实现了在10~50 kW功率范围多个功率水平的稳定长时间连续运行,典型运行结果为16 kW/3000 s、26 kW/900 s、46 kW/1800 s、50 kW/300 s,特别在输出功率32 kW连续稳定工作了400 min。这是国内首次研制出小时级连续工作的中等功率回旋管。

多示例学习的簇频繁性分析及双角度融合嵌入

多示例学习(Multi-Instance Learning,MIL)的训练数据是由若干个未带标记的示例组成的带标记的包,基于嵌入的方法,通过将包嵌入成单向量来解决包表示问题,然而大部分现有方法忽略了示例与包的联系,难以保证所选示例的代表性.同时,单角度的嵌入方法无法有效地提取正、负包的差异信息,使嵌入向量的质量较差.提出一种多示例学习的簇频繁性分析及双角度融合嵌入(FADE).簇频繁性分析技术从正、负子空间中分别筛选部分示例作为子空间的簇心,依据簇心将子空间聚类成簇,再计算簇频繁性指标,选择频繁性较高的簇的簇心组成子空间代表示例集.双角度融合嵌入技术基于正、负子空间代表示例集和差值嵌入函数,分别从正、负角度挖掘信息,融合两个角度信息获得最终的嵌入向量.在29个数据集上与七个MIL算法进行了对比实验,结果表明,FADE的分类准确率总体上优于七个对比算法,在图像数据集上有显著优势,在文本和网页数据集上也表现良好.

基于相对熵权重组合的多阶段贮存数据融合评估方法

目的针对目前仅使用加速贮存数据获取产品性能参数退化信息导致贮存寿命评估可信度低的问题,融合自然贮存数据和加速贮存数据2阶段的数据结果。方法采用基于最小二乘和相对熵的方法,确定不同阶段数据的权重比例,最后形成融合2阶段数据的评估方法,并在扭杆和控制放大器2个产品上进行自然贮存数据与加速贮存数据的融合分析。结果实现了2个产品贮存寿命的高置信度评估,扭杆自然贮存数据和加速贮存数据的权重比例分别为0.263和0.737,控制放大器自然贮存数据和加速贮存数据的权重比例分别为0.776和0.224。结论融合评价的方法能够结合2部分数据的退化特点,给出一个更有信服力的评价结果。

基于多源信息融合和WOA-CNN-LSTM的外脚手架隐患分类预警研究

面对施工现场外脚手架隐患信息的多样性,传统的基于传感器监测的单一信号预警研究存在容错力不佳、含有信息有限等问题。针对施工现场外脚手架“图像+监测”数据,提出一种基于数据层和特征层信息融合的脚手架隐患分类预警方法。首先,利用Revit三维建模软件建立外脚手架实体模型,对不同初始隐患下的外脚手架进行有限元分析,划分隐患预警等级;其次,利用无迹卡尔曼滤波算法(Unscented Kalman Filter,UKF)及卷积长短时记忆网络(Convolutional Neural Network-Long Short Term Memory Network,CNN-LSTM)实现脚手架同类信息数据层融合及异类信息特征层融合;最后,通过实时收集西安市某在建项目落地式双排扣件式钢管脚手架隐患信息,对其进行分类预警,并使用鲸鱼优化算法(Whale Optimization Algorithm,WOA)对CNN-LSTM网络进行参数优化,发现隐藏节点个数为30、学习率为0.0072、正则化系数为1×10^(-4)时分类效果最佳,优化后预警精度达到了91.4526%。通过可视化WOA-CNN-LSTM、CNN-LSTM、CNN-SVM(Support Vector Machine,支持向量机)及CNN-GRU(Gate Recurrent Unit,门控循环单元)分类预警结果,证实了优化后的CNN-LSTM网络在脚手架分类预警方面的优越性。

多源数据融合的深埋隧道岩爆预测方法

为提高隧道设计阶段未开挖区域的岩爆预测准确性,提出了一种多源数据融合的隧道岩爆预测方法.综合隧道勘察和施工阶段的不同地质信息,采用基于证据关联系数的加权融合技术,构建了隧道精细化三维动态地质模型.建立了基于强度理论的岩爆可能性判定方法和基于能量理论的岩爆烈度预测方法,通过Hoek-Brown强度准则判断围岩是否发生岩爆,利用储能极限阈值和能量释放指数划分岩爆烈度,并将其应用于四川某隧道工程中.结果表明,所提方法可以实现隧道掌子面前方30 m范围内的岩爆精准预测,预测结果与隧道开挖实际岩爆的吻合率为95.8%.因此,该预测方法能够在隧道施工前预判岩爆烈度,为隧道岩爆防治提供指导.