基于图像处理的公交车内人群异常情况检测

为有效地监控公交车这一特定环境中人群的异常行为,提出一种公交车内人群异常情况检测的方法。对视频图像确立感兴趣区域,进行预处理;通过改进Vi Be算法提取运动目标,引入多尺度滑窗算法确定识别区域;结合连续多帧识别区域进行改进卷积神经网络算法的异常行为识别,通过识别结果判断公交车内人群是否异常。与传统方法的比较结果表明,该算法的检测正确率较高,可达93.5%,误检率较低,仅为1.6%,在实际应用中具有较高的参考价值。     

人群疏散平移矩阵算法

建筑物的通行能力研究近年来受到广泛关注,安全高效地流通人群是专家学者研究的重点。首先,通过一组节点和边的定义将建筑物表示为网络,进而利用Dijkstra最短路算法生成最短路径集合,最后在该集合的基础上提出平移矩阵算法,利用0-1矩阵的更新模拟人群在建筑路径上的流动过程。通过设置多组教学楼疏散场景并与Pathfinder软件的对比来验证平移矩阵算法的有效性。对比结果显示,不同密度下两种方法的总疏散时间差都约为50 s,而运算速度平移矩阵算法比Pathfinder软件快10~110倍。该算法为人群疏散的快速模拟提供了新思路,为疏散策略的实时生成提供了可能性。     

Long-Time gap crowd prediction with a Two-Stage optimized spatiotemporal Hybrid-GCGRU

Crowd prediction is a crucial aspect of modern society, facilitating numerous decision-making processes, such as hazard detection and facility maintenance. Conventional crowd prediction relies on rule-based models which involve tedious formation steps and are error-prone. Hence, this study adopts spatiotemporal deep learning models for crowd prediction in order to exploits both temporal and spatial features of the dataset. There are three major limitations in previous studies: firstly, the prediction length of the time series models is relatively short (i.e., limited to single-digit time steps); Secondly, most studies using time series prediction do not consider the features in the neighboring locations other than the target location; Finally, only the temporal pattern on regular days is exploited, whilst the data variations during holidays can greatly reduce the prediction accuracy. Therefore, a particle swarm optimized (PSO) Hybrid-Graph Convolutional Gated Recurrent Unit (HGCGRU) model, entitled PSO-HGCGRU, is proposed to increase the prediction accuracy in various situations. The PSO-HGCGRU model is composed of two parts, 1) a hybrid model structure, which allows the model adapting to both regular and holiday features, and 2) the GCGRU model, in which the graph convolutional layer to extract the spatial features and the GRU layers to extract the temporal dependencies, to undertake spatiotemporal prediction. The Two-Stage Long-Time Gap Prediction method is applied to the proposed PSO-HGCGRU model with a newly designed two-stage mechanism to optimize the length of the time gap. Results show that the proposed model generally outperforms the baseline models by around 40% in accuracy.     

基于卷积神经网络的实时人群密度估计

针对传统实时人群密度估计方法存在误差大、分类效果不佳等缺陷,提出了基于 卷积神经网络的实时人群密度估计方法。通过对比4 种常见网络结构：AlexNet、VGGNet、 GoogLeNet 和ResNet 的准确度与实时性,选择综合性较好的GoogLeNet 作为人群密度估计的 模型,利用关键帧截取技术实现人群密度的实时估计并简要分析人群密度特征图。最后用实例 验证了该方法的实时性与准确度,证明了其可行性。     

Measurement and modelling of the interfacial tension of triglyceride mixtures in contact with dense gases

The interfacial tension of various triglyceride mixtures in contact with carbon dioxide is compared. The interfacial tension of these triglyceride mixtures is similar. The interfacial tension in contact with carbon dioxide and nitrogen decreases with increasing pressure. Concerning the interfacial tension in contact with carbon dioxide, the interfacial tension increases at higher pressure with rising temperature. The interfacial tension of triglyceride mixtures and of many other liquids in contact with carbon dioxide can be described as a function of the carbon dioxide density.The modelling of the interfacial tension of triglycerides in contact with carbon dioxide or nitrogen using the density gradient theory (Macromolecules 14 (1981) 361) and the Sanchez-Lacombe equation of state (J. Phys. Chem. 80 (1976a/b) 2352 and 2568) is successful if one parameter for nitrogen and two parameters for carbon dioxide are employed. Nevertheless, an appropriate equation of state must be utilized for modelling the interfacial tension of a system. For complex systems, e.g. for triglyceride mixtures, it is indispensable to check the equation of state by means of experimental equilibrium data (e.g. density, volume, or mole fraction).     

基于观点认同的群体情绪模型

为了能够有效地模拟群体环境中群体情绪的演化过程, 以情绪感染为理论基础, 提出一种新的基于观点认同的群体情绪模型。该模型引入个体观点值、情绪触发阈值来实现观点交互与认同对群体情绪演化所产生的影响作用。观点交互采用有界信任Deffuant模型来实现。通过模拟发现, 观点交互不仅加速了群体情绪融合过程, 而且在多观点群体环境中对群体情绪融合态势也具有决定性的作用。实验模拟结果表明, 本模型不仅能够体现情绪感染过程中个体的无意识、自动感染过程, 而且也能够体现个体对传染对象的选择性过程, 弥补了情绪感染过程中个体无认知的缺陷。     

基于GPU的大规模人群疏散模拟*

基于中央处理器(CPU)串行的人群疏散传统方法对于人群规模较少的场景,可以得到良好的疏散模拟效果,但在人群密度较高的场景中,难以达到实时模拟的要求。为了克服上述问题,实现了一种基于图形处理器(GPU)的人群疏散模拟的方法。该方法通过对个体寻径算法的优化,不仅能使个体快速准确地智能寻径,而且将个体寻径独立性与图形处理器高性能计算特性进行结合,充分利用了图形处理器强大的并行计算能力,从而大幅度提高了人群疏散模拟的人群规模,使人群疏散模拟的实时性得到增强。     

随机人群运动荷载作用下大跨度连廊的振动响应

基于社会力模型,研究了随机人群运动荷载作用下大跨连廊结构的振动舒适度问题。通过结合交通学领域、生物力学领域对人群运动特性及人体运动参数的统计分析,对随机人群运动进行仿真,得到不同密度工况下,任意行人在任意时刻的步速、落点位置等参数,建立随机人群荷载模型,计算不同工况密度下结构的加速度峰值。研究表明,加速度峰值与行人步速关系很大,随着行人密度增大,行人速度减小,结构加速度响应则先增大后减小。根据得到的结构加速度峰值与行人密度关系曲线,可以找出出现结构振动舒适度问题的行人密度区间。     

A computational and experimental study of ultra fine water mist as a total flooding agent

Computational fluid dynamics (CFD) calculations were carried out to design total flooding fire tests in a 28 m³ compartment for an ultra fine water mist (<10 μm). The exit momentum of the mist produced by a proprietary ultrasonic generator technology was extremely low with a mist discharge velocity below 1 m/s. The mist was discharged with multiple floor outlets equally spaced around the centrally located 120 kW pool-like gas fire. The transport of mist and its interaction with the fire was simulated by Fluent, a commercial CFD model. Lagrangian Discrete Phase Model (DPM) was used for droplets. Simulation predicted extinguishment within 10 s with a mist delivery rate of 1 l/min. However, in total flooding fire tests conducted, extinction times were more than 5 min. Additional computations approximating the ultra fine mist (UFM) as a dense gas agreed well with the observed transport timescales of minutes indicating that UFM behaves like a gas. Further, the mist–fire interaction needs a multi-phase Euler–Euler approach with a droplet vaporization model.