近红外光谱的北方寒地土壤含水率预测模型研究

我国北方寒地温差大,土壤温差对近红外光谱测量土壤墒情有较大影响。针对这一问题,以北方寒地土壤为研究对象,探究大范围温度胁迫下（-20～40 ℃）土壤的近红外光谱与土壤不同含水率之间的关系预测模型方法。选取黑龙江八一农垦大学农学院试验基地中的黑土,经烘干、过筛等操作处理后配置含水率范围在15%～50%内八种不同湿度的土壤样品,建立北方寒地土壤大范围温度胁迫下土壤的近红外光谱信息与含水率之间的定量预测模型。在全波段光谱数据的基础上,结合五种不同光谱信号预处理方法,采用BP神经网络算法、优化支持向量机算法（SVM）、高斯过程算法（GP）三种智能算法建立北方寒地土壤近红外光谱与含水率的预测模型并验证模型的效果。利用69组数据进行训练建模, BP神经网络相关参数设置为学习速率0.05,最大训练次数设置为5 000,隐层单元数确定为20;SVM采用径向基函数,并利用leave-one-out cross validation确定了最佳惩罚参数为0.87,使模型预测的准确性提高;高斯过程算法内部采用马顿核。模型的定量评估采用决定系数(R2)和均方根误差（RMSE）。结果表明,在建立的全部BP神经网络模型中,效果最佳的为S_G-BP神经网络模型,模型的R2为0.960 9,RMSE为2.379 7;在SVM模型中SNV-SVM模型的效果最好,模型的R2为0.991 1,RMSE为1.081 5;在GP模型中S_G-GP模型的效果最好,模型的R2为0.928,RMSE为3.258 1,综上基于SNV预处理的SVM模型训练效果最优。利用剩余的35组光谱数据作为预测集验证模型性能,经模型对比分析发现基于SVM算法的预测模型效果优于其他两种算法,其中基于S_G的SVM模型效果最优,其预测模型的R2和差RMSE分别为0.992 1和0.736 9。综合建模集与预测集的参数最终确定基于S_G的SVM模型为最佳模型。此模型可以作为大范围温度胁迫条件下（寒地）的土壤含水率有效预测方法,为设计优化适宜寒地便携式近红外土壤含水率快速测量仪提供科学依据。     

全透射近红外光谱的苹果整果糖度在线检测模型优化

光谱质量、样本个体差异、检测系统和建模算法等多种因素共同决定水果糖度检测模型的预测精度和稳定性。采用自主研发的短积分全透射近红外在线检测系统以5 ms积分时间和0.5 m·s-1运行速度在线获取了“富士”苹果全透射光谱信号。不同姿态获取的透射光谱强度差异明显,但曲线走势相近,均在920 nm波段具有最大的光谱强度,在850 nm波段存在波谷。采用移动平均平滑、标准正态变量变换和多元散射校正等预处理方法有效去除原始光谱的随机噪声和基线偏差,减小了样本检测姿态引起的光谱差异。为分析不同检测姿态对苹果整果糖度预测模型的影响,构建了单一姿态局限模型和多姿态通用模型,结果表明基于全位点平均透射光谱构建的单一姿态局限模型对检测姿态具有很大的局限性,而多姿态通用模型预测能力较单一检测姿态相当,但却对不同的检测姿态具有更强的适用能力。为进一步提高光谱信号质量,优化模型预测能力,采用信号强度阈值优选方法实现了苹果整果糖度预测模型优化,发现移除中央位点获取的透射光谱信号,有利于提高苹果整果糖度预测模型精度。多姿态通用信号强度优化模型综合考虑不同姿态获取的光谱信息有效性,有效提升了通用信号强度优化模型的预测能力和稳定性,当多姿态通用模型中信号强度阈值为5 000时,模型预测性能最佳,其预测参数R_p,RMSEP和RPD分别为0.79,0.84%和1.58。表明短积分全透射近红外在线检测系统用于不同姿态苹果糖度预测是可行的,多姿态通用模型的建立,扩大了模型在不同姿态的预测稳健性,短积分光谱采集方式结合信号强度阈值优选方法提升了光谱信号的质量和模型的预测能力。     

深度神经网络和高光谱显微图像的二维材料纳米片识别

近年来,二维材料由于其独特的性质而受到了广泛关注。在制备二维层状晶体的各种方法中,机械剥离法获得的薄层二维材料晶体质量高,适用于基础研究及性能演示。然而用机械剥离法从衬底上获得的材料具有一定的随机性,可能包含了少许相对较厚的部分。实现对这些二维薄层材料有效、快速且智能化的表征有利于促进二维材料性能的进一步研究。提出了一种基于深度学习的表征方法,通过搭建的编解码结构的卷积神经网络语义分割算法,可以根据光学显微镜图像进行分割和快速识别二维材料纳米片。卷积神经网络作为深度学习在图像处理领域中的典型算法,能够对光学显微镜图像中的复杂信息进行特征提取。首先采用机械剥离制备MoS₂纳米片样本,通过光学显微镜采集高光谱图像并对样本进行标记,根据样本的厚度范围标记出不同的区域,对标记后的图像进一步处理,包括图像的颜色校准和剪切操作,得到用于网络训练和测试的数据集。针对光学图像中二维纳米薄片存在的低对比度、碎裂等特点,编码时加入残差结构和金字塔池化模型,有助于特征信息的提取;解码时融合编码路径中提取的浅层特征信息,以提高网络分割精度。实验中采用带权重的交叉熵损失函数解决类别数量不平衡问题和采用数据增强扩大数据集。对训练后的网络测试结果表明,模型像素精度为97.38%,平均像素精度为90.38%,均交并比为75.86%。之后通过迁移学习成功地对剥离的单层和双层石墨烯纳米片样本进行了识别,均交并比达到了81.63%,表明该方法具有普适性。通过MoS₂和石墨烯纳米片的识别演示,实现了深度学习在二维材料的光学显微镜图像中的成功应用。该方法有望在更多的二维材料上得到扩展并突破自动动态处理光学显微镜图像的问题,同时为其他纳米材料的高光谱图像处理提供参考。     

In situ measurement on nonuniform velocity distributionin external detonation exhaust flow by analysis ofspectrum features using TDLAS

Instantaneous and precise velocity sensing is a critical part of research on detonation mechanism and flow evolution.This paper presents a novel multi-projection tunable diode laser absorption spectroscopy solution,to provide a real-time and reliable measurement of velocity distribution in detonation exhaust flow with obvious nonuniformity.Relations are established between overlapped spectrums along probing beams and Gauss velocity distribution phantom according to the frequency shifts and tiny variations in components of light-of-sight absorbance profiles at low frequencies analyzed by the fast Fourier transform.With simulated optical measurement using H2O feature at 7185.6 cm-1 carried out on a phantom generated using a simulation of two-phase detonation by a two-fluid model,this method demonstrates a satisfying performance on recovery of velocity distribution profiles in supersonic flow even with a noise equivalent absorbance up to 2×10^-3.This method is applied to the analysis of rapidly decreasing velocity during a complete working cycle in the external flow field of an air-gasoline detonation tube operating at 25 Hz,and results show the velocity in the core flow field would be much larger than the arithmetic average from traditional tunable diode laser doppler velocimetry.This proposed velocity distribution sensor would reconstruct nonuniform velocity distribution of high-speed flow in low cost and simple operations,which broadens the possibility for applications in research on the formation and propagation of external flow filed of detonation tube.     

Modeling and character analyzing of multiple fractional-order memcapacitors in parallel connection

Recently,the memory elements-based circuits have been addressed frequently in the nonlinear circuit theory due to their unique behaviors.Thus,the modeling and characterizing of the mem-elements become essential.In this paper,the analysis of the multiple fractional-order voltage-controlled memcapacitors model in parallel connection is studied.Firstly,two fractional-order memcapacitors are connected in parallel,the equivalent model is derived,and the characteristic of the equivalent memcapacitor is analyzed in positive or negative connection.Then a new understanding manner according to different rate factor K and fractional orderαis derived to explain the equivalent modeling structure conveniently.Additionally,the negative order appears,which is a consequence of the combination of memcapacitors in different directions.Meanwhile,the equivalent parallel memcapacitance has been drawn to determine that multiple fractional-order memcapacitors could be calculated as one composite memcapacitor.Thus,an arbitrary fractional-order equivalent memcapacitor could be constructed by multiple fractional-order memcapacitors.     

Influence of various shapes of nanoparticles on unsteady stagnation-point flow of Cu-H2O nanofluid on a flat surface in a porous medium: A stability analysis

The nanofluid and porous medium together are able to fulfill the requirement of high cooling rate in many engineering problems. So, here the impact of various shapes of nanoparticles on unsteady stagnation-point flow of Cu-H₂O nanofluid on a flat surface in a porous medium is examined. Moreover, the thermal radiation and viscous dissipation effects are considered. The problem governing partial differential equations are converted into self-similar coupled ordinary differential equations and those are numerically solved by the shooting method. The computed results can reveal many vital findings of practical importance. Firstly, dual solutions exist for decelerating unsteady flow and for accelerating unsteady and steady flows, the solution is unique. The presence of nanoparticles affects the existence of dual solution in decelerating unsteady flow only when the medium of the flow is a porous medium. But different shapes of nanoparticles are not disturbing the dual solution existence range, though it has a considerable impact on thermal conductivity of the mixture. Different shapes of nanoparticles act differently to enhance the heat transfer characteristics of the base fluid, i.e., the water here. On the other hand, the existence range of dual solutions becomes wider for a larger permeability parameter related to the porous medium. Regarding the cooling rate of the heated surface, it rises with the permeability parameter, shape factor (related to various shapes of Cu-nanoparticles), and radiation parameter. The surface drag force becomes stronger with the permeability parameter. Also, with growing values of nanoparticle volume fraction, the boundary layer thickness (BLT) increases and the thermal BLT becomes thicker with larger values of shape factor. For decelerating unsteady flow, the nanofluid velocity rises with permeability parameter in the case of upper branch solution and an opposite trend for the lower branch is witnessed. The thermal BLT is thicker with radiation parameter. Due to the existence of dual solutions, a linear stability analysis is made and it is concluded that the upper branch and unique solutions are stable solutions.     

Traffic flow prediction based on BILSTM model and data denoising scheme

Accurate prediction of road traffic flow is a significant part in the intelligent transportation systems. Accurate prediction can alleviate traffic congestion, and reduce environmental pollution. For the management department, it can make effective use of road resources. For individuals, it can help people plan their own travel paths, avoid congestion, and save time. Owing to complex factors on the road, such as damage to the detector and disturbances from environment, the measured traffic volume can contain noise. Reducing the influence of noise on traffic flow prediction is a piece of very important work. Therefore, in this paper we propose a combination algorithm of denoising and BILSTM to effectively improve the performance of traffic flow prediction. At the same time, three denoising algorithms are compared to find the best combination mode. In this paper, the wavelet (WL) denoising scheme, the empirical mode decomposition (EMD) denoising scheme, and the ensemble empirical mode decomposition (EEMD) denoising scheme are all introduced to suppress outliers in traffic flow data. In addition, we combine the denoising schemes with bidirectional long short-term memory (BILSTM) network to predict the traffic flow. The data in this paper are cited from performance measurement system (PeMS). We choose three kinds of road data (mainline, off ramp, on ramp) to predict traffic flow. The results for mainline show that data denoising can improve prediction accuracy. Moreover, prediction accuracy of BILSTM+EEMD scheme is the highest in the three methods (BILSTM+WL, BILSTM+EMD, BILSTM+EEMD). The results for off ramp and on ramp show the same performance as the results for mainline. It is indicated that this model is suitable for different road sections and long-term prediction.     

An extended improved global structure model for influential node identification in complex networks

Accurate identification of influential nodes facilitates the control of rumor propagation and interrupts the spread of computer viruses. Many classical approaches have been proposed by researchers regarding different aspects. To explore the impact of location information in depth, this paper proposes an improved global structure model to characterize the influence of nodes. The method considers both the node's self-information and the role of the location information of neighboring nodes. First, degree centrality of each node is calculated, and then degree value of each node is used to represent self-influence, and degree values of the neighbor layer nodes are divided by the power of the path length, which is path attenuation used to represent global influence. Finally, an extended improved global structure model that considers the nearest neighbor information after combining self-influence and global influence is proposed to identify influential nodes. In this paper, the propagation process of a real network is obtained by simulation with the SIR model, and the effectiveness of the proposed method is verified from two aspects of discrimination and accuracy. The experimental results show that the proposed method is more accurate in identifying influential nodes than other comparative methods with multiple networks.     

Passenger management strategy and evacuation in subway station under Covid-19

Under the background of Covid-19 sweeping the world, safe and reasonable passenger flow management strategy in subway stations is an effective means to prevent the spread of virus. Based on the social force model and the minimum cost model, the movement and path selection behavior of passengers in the subway station are modeled, and a strategy for passenger flow management to maintain a safe social distance is put forward. Take Qingdao Jinggangshan Road subway station of China as the simulation scene, the validity of the simulation model is verified by comparing the measured value and simulation value of the time required for passengers from getting off the train to the ticket gate. Simulation results indicate that controlling the time interval between incoming passengers at the entrance can effectively control the social distance between passengers and reduce the risk of epidemic infection. By comparing the evacuation process of passengers under different initial densities, it is found that the greater the initial density of passengers is, the longer the passengers are at risk social distance. In the process of passenger emergency evacuation, the stairs/escalators and ticket gates are bottleneck areas with high concentration of passenger density, which should be strictly disinfected many times on the basis of strictly checking the health code of incoming passengers and controlling the arrival time interval. The simulation results of this paper verify the harmfulness of passenger emergency evacuation without protective measures, and provide theoretical support for the operation and management of subway station under the epidemic situation.     

Effect of astrocyte on synchronization of thermosensitive neuron-astrocyte minimum system

Astrocytes have a regulatory function on the central nervous system (CNS), especially in the temperature-sensitive hippocampal region. In order to explore the thermosensitive dynamic mechanism of astrocytes in the CNS, we establish a neuron-astrocyte minimum system to analyze the synchronization change characteristics based on the Hodgkin-Huxley model, in which a pyramidal cell and an interneuron are connected by an astrocyte. The temperature range is set as 0 ℃-40 ℃ to juggle between theoretical calculation and the reality of a brain environment. It is shown that the synchronization of thermosensitive neurons exhibits nonlinear behavior with changes in astrocyte parameters. At a temperature range of 0 ℃-18 ℃, the effects of the astrocyte can provide a tremendous influence on neurons in synchronization. We find the existence of a value for inositol triphosphate (IP₃) production rate and feedback intensities of astrocytes to neurons, which can ensure the weak synchronization of two neurons. In addition, it is revealed that the regulation of astrocytes to pyramidal cells is more sensitive than that to interneurons. Finally, it is shown that the synchronization and phase transition of neurons depend on the change in Ca²⁺ concentration at the temperature of weak synchronization. The results in this paper provide some enlightenment on the mechanism of cognitive dysfunction and neurological disorders with astrocytes.