Configuration of more than N DWDM channels with only one N × N cyclic-AWG-based wavelength routing device

A wavelength routing device based on only one N × N cyclic arrayed waveguide grating (AWG) having easy extended channels configuration is presented in this paper. It is easy to extend the dense wavelength division multiplexing (DWDM) channel configuration through this wavelength routing device. According to the cyclic wavelength of AWG, the wavelength routing devices are easy to configure more than N extended DWDM channels through cascading more proper tunable fiber Bragg gratings (FBGs). With only one 8 × 8 AWG, two different wavelength routing structures were built to evaluate static crosstalk and the bit-error-rate (BER). Three of the 16 inputted DWDM channels were demonstrated to verify that the proposed wavelength routing device, with only one 8 × 8 AWG, could configure extended DWDM channels without interfering with other channels. The results show that the wavelength routing device can produces a better performance and offers a cheaper way to extend the DWDM channel configuration for a dynamic network.     

CO‐Free Hydrogen Production by Ethanol Steam Reforming in a Pd–Ag Membrane Reactor

In this work, the ethanol steam reforming (ESR) reaction has been studied by using a dense Pd–Ag membrane reactor (MR) by varying the water/ethanol molar ratio between 3:1 and 9:1 in a temperature range of 300–400 °C and at 1.3 bar as reaction pressure. The MR was packed with a commercial Ru‐based catalyst and a constant sweep gas flow rate in counter current mode was used. The influence of the temperature and the feed molar ratio on different parameters such as the ethanol conversion, the hydrogen production, the hydrogen yield and the CO‐free hydrogen recovery has been evaluated.     

Reference guided image super-resolution via efficient dense warping and adaptive fusion

Due to the limited improvement of single-image based super-resolution (SR) methods in recent years, the reference based image SR (RefSR) methods, which super-resolve the low-resolution (LR) input with the guidance of similar high-resolution (HR) reference images are emerging. There are two main challenges in RefSR, i.e. reference image warping and exploring the guidance information from the warped references. For reference warping, we propose an efficient dense warping method to deal with large displacements, which is much faster than traditional patch (or texture) matching strategy. For the SR process, since different reference images complement each other, and have different similarities with the LR image, we further propose a similarity based feature fusion strategy to take advantage of the most similar reference regions. The SR process is realized by an encoder–decoder network and trained with pixel-level reconstruction loss, degradation loss and feature-level perceptual loss. Extensive experiments on three benchmark datasets demonstrate that the proposed method outperforms state-of-the-art SR methods in both subjective and objective measurements.     

Exploring the biomechanics of walking and crowd “flow”

Crowd movement simulation models are generally based on aggregated speed and flow data collected more than 50 years ago. There appears to be no validated modelling capability to include the impact of recent and future changes in population demographics, resulting from an ageing population and increasing obesity rates. New analytical approaches and data gathering are required to successfully model crowd movement and safety for current and future generations. This study carried out (a) a review of the primary components of crowd movement, demographics and analytical techniques, (b) prototype experiments to investigate age-related aspects of space and potential points of contact and (c) a new predictive model for crowd flow analysis based on pedestrian biomechanics and anthropometric data. The model uses the physical space taken up by the biomechanical walking process and the spatial buffer between points of potential contact with other pedestrians to predict the speed of movement at different levels of congestion. The new analytical model was used to predict single file speeds (for people with different demographics in congested space), which compared well with published experimental data. The next steps for model development for wider “flows” and additional experiments to provide data sets for wider demographics are also proposed.     

基于多尺度多任务卷积神经网络的人群计数

在智能监控领域，实现人群计数具有重要价值，针对人群尺度不一、人群密度分布不均及遮挡等问题，提出一种多尺度多任务卷积神经网络（MMCNN）进行人群计数的方法。首先提出一种新颖的自适应人形核生成密度图描述人群信息，消除人群遮挡影响；其次通过构建多尺度卷积神经网络解决人群尺度不一问题，以多任务学习机制同时估计密度图及人群密度等级，解决人群分布不均问题；最后设计一种加权损失函数，提高人群计数准确率。在UCF_CC_50和World Expo'10数据库上进行了评估，验证了自适应人形核的有效性。实验结果表明：所提算法比Sindagi等的方法（SINDAGI V A，PATEL V M.CNN-based cascaded multi-task learning of high-level prior and density estimation for crowd counting.Proceedings of the 2017 14th IEEE International Conference on Advanced Video and Signal Based Surveillance.Piscataway，NJ：IEEE，2017：1-6）在UCF_CC_50数据库上平均绝对误差（MAE）数值和均方误差（MSE）数值分别降低约1.7和45；与Zhang等的方法（ZHANG Y，ZHOU D，CHEN S，et al.Single-image crowd counting via multi-column convolutional neural network.Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition.Washington，DC：IEEE Computer Society，2016：589-597）相比，在World Expo'10数据库上所提算法的MAE值降低约1.5，且在真实公共汽车数据库上仅0~3人的计数误差，表明其实用性较强。     

Densely connected convolutional network block based autoencoder for panorama map compression

As a novel virtual reality (VR) format, panorama maps are attracting increasing attention, while the compression of panorama images is still a concern. In this paper, a densely connected convolutional network block (dense block) based autoencoder is proposed to compress panorama maps. In the proposed autoencoder, dense blocks are specially designed to reuse feature maps and reduce redundancy of features. Meanwhile, a loss function, which imports a position-dependent weight item for each pixel, is proposed to train and adjust network parameters, in order to make the autoencoder fit to properties of panorama maps. Based on the proposed autoencoder and the weighted loss function, a greedy block-wise training scheme is also designed to avoid gradient vanishing problem and speed up training. During training process, the autoencoder is divided into several sub-nets. After each sub-net is trained separately, the whole network is fine-tuned to achieve the best performance. Experimental results demonstrate that the proposed autoencoder, compared with JPEG, saves up to 79.69 % bit rates, and obtains 7.27dB gain in BD-WS-PSNR or 0.0789 gain in BD-WS-SSIM. The proposed autoencoder also outperforms JPEG 2000, HEVC and VVC in both BD-WS-PSNR and BD-WS-SSIM. Meanwhile, subjective results show that the proposed autoencoder can recover details of panorama images, and reconstruct maps with high visual quality.     

JYX高浓度黏稠物料加压旋转流变仪的研制

鉴于高浓度黏稠物料(如原生煤泥、城市污泥、造纸污泥等一类工业副产品或固体废弃物)若无添加剂作用在常压下是一种分散结团物质，流变参数测定过程中常出现被测介质“打滑”和容易“离析”等问题，研制出一种加压旋转流变仪(压力最高可达30MPa)，测量时将物料封闭加压，从而达到此类介质加压测量流变特性的目的。     

Analysis of crowds' movement using Twitter

Over the last decade, the infrastructure supporting the smart city has lived together with and was surpassed by the rise of social media. The tremendous growth of both mobile devices and social media users has unearthed a new kind of services in the so‐called location‐based social networks (LBSNs). In this new scenario, the term crowdsensing refers to sharing data collected by sensing humans with the aim of measuring phenomena of common interest. Crowd‐sourced location data provide the ability to study, for the first time, the movement of individuals in urban environments. In this paper, we address the problem of monitoring crowds, whereabouts and movement, which can assist decision making in education, emergency training, urban planning, traffic engineering, etc. Precisely, two‐phase density‐based analysis for collectives and crowds (2PD‐CC) is a novel methodology over public data in LBSN, which combines density‐based clustering, outlier detection a topic modeling over a region under study to detect, predict, and explain abnormal group behavior. In order to validate the methodology and its potential application to full‐scale problems, an experiment over Twitter data was performed in Madrid city.     

基于多级特征和混合注意力机制的室内人群检测网络

针对室内人群目标尺度和姿态多样性、人头目标易与周围物体特征混淆的问题,提出了一种基于多级特征和混合注意力机制的室内人群检测网络（MFANet）。该网络结构包括三部分,即特征融合模块、多尺度空洞卷积金字塔特征分解模块以及混合注意力模块。首先,通过将浅层特征和中间层特征信息融合,形成包含上下文信息的融合特征,用于解决浅层特征图中小目标语义信息不丰富、分类能力弱的问题;然后,利用空洞卷积增大感受野而不增加参数的特性,对融合特征进行多尺度分解,形成新的小目标检测分支,实现网络对多尺度目标的定位和检测;最后,用局部混合注意力模块来融合全局像素关联空间注意力和通道注意力,增强对关键信息贡献大的特征,来增强网络对目标和背景的区分能力。实验结果表明,所提方法在室内监控场景数据集SCUT-HEAD上达到了0.94的准确率、0.91的召回率和0.92的F1分数,在召回率、准确率和F1指标上均明显优于当前用于室内人群检测的其他算法。