Multitask bilateral learning for real‐time image enhancement

Nowadays, deep neural networks (DNNs) for image processing are becoming more complex; thus, reducing computational cost is increasingly important. This study highlights the construction of a DNN for real‐time image processing, training various image processing operators efficiently through multitask learning. For real‐time image processing, the proposed algorithm takes a joint upsampling approach through bilateral guided upsampling. For multitask learning, the overall network is based on an encoder‐decoder architecture, which consists of encoding, processing, and decoding components, in which the encoding and decoding components are shared by all the image processing operators. In the processing component, a semantic guidance map, which contains processing information for each image processing operator, is estimated using simple linear shifts of the shared deep features. Through these components, the proposed algorithm requires an increase of only 5% in the number of parameters to add another image processing operator and achieves faster and higher performance than that of deep‐learning‐based joint upsampling methods in local image processing as well as global image processing.     

基于字符结构知识的车牌汉字快速识别技术

车牌汉字识别效率一直是制约车牌自动识别系统推广应用的瓶颈。采用多级分类方法,充分利用车牌汉字图像的空问结构特征(像素空间分布投影,字符的笔段类型、数量及拓扑关系等)可实现车牌汉字的快速辨识。实践表明,该方法能快速从车牌汉宁图像中辨识出汉字字符,识别准确度可达98％。     

一种基于GOP的MPEG-2媒体流动切割与合并方法

MPEG-2视频编码技术由于其较高的压缩比和高质量图像，在数字电视及电影等方面具有重要应用。由于编码算法相对较为复杂，迫切需要解决MPEG-2媒体流的切割和合并问题，从而实现有效的节目制作。本文正是针对MPEG-2媒体流，提出了一种基于GOP的压缩域视频分割与合并方法。通过分析媒体流的数据结构，详细蛤出了算法的原理和实现过程。最后，还通过实验证明了该方法的可行性和高效率。     

Broadband MIMO Antenna System for 5G Operations in Mobile Phones

A broadband 8‐antenna multiple‐input and multiple‐output (MIMO) system working for 5G new radio N77 (3.3‐4.2 GHz) in mobile terminals is proposed. In particular, each antenna element/unit consists of a monopole‐type meandered branch and an inverted U‐shaped self‐isolated antenna branch; and the antenna unit is fed by the meandered branch. The antenna unit has two resonances generated separately by each branch of the antenna unit. Since the monopole couples strongly with the self‐isolated antenna element that has the inherent self‐isolation characteristics, quite good isolation for the broadband 8‐antenna MIMO system can be obtained. The results such as antenna efficiency, radiation patterns, envelope correlation coefficient, and channel capacity are also presented.     

面向多用户软件系统的可靠性分配的故障树分析法

本文首先对目前已有的各种分配技术与方法的特点进行了探讨,介绍了如何采用故障树技术求得导致系统失效的功能模块的最小割集,首次提出运用故障树技术进行各模块可靠性的划分和确定,并提出了一种新的算法来快速、有效地确定各模块的重要度,结合考虑软件的总体开发费用,创建了一个新的基于软件实用性的可靠性分配模型并详细讨论了其具体运用,最后本文对该方法可能的改进进行了一些探讨。     

一种新的海量数据分类方法

使用支持向量机对非线性可分数据进行分类的基本思想是将样本集映射到一个高维线性空间使其线性可分。文章则基于Jordan曲线定理,提出了一种通用的基于分类超曲面的分类法,它是通过直接构造分类超曲面,根据样本点关于分类曲面的围绕数的奇偶性进行分类的一种新分类判断算法,不需作升维变换,不需要考虑使用何种核函数,而直接地解决非线性分类问题。对数据分类应用的结果说明:基于分类超曲面的分类法可以有效地解决非线性数据的分类问题,并能够提高分类效率和准确度。     

遥感影像多尺度分割中最优尺度的选取及评价

多尺度分割是面向对象图像分析技术的前提和关键,多尺度分割的质量直接影响着面向对象分类的精度,但尺度选择仍然是多尺度分割中的一个难题。针对此问题,根据遥感影像的最优分割尺度与影像上目标复杂度密切相关的事实,提出了一种自上而下基于分割对象复杂度选取最优尺度的方法。该方法在分割过程中,提取每一对象的影像特征构建其复杂度函数,通过设置阈值,经迭代计算来确定每一对象的最优分割尺度,进而得到具有全局最优尺度的分割结果,并将其应用于ZY-3多光谱数据和GF-2融合影像,得到分割和分类结果。并将其与单一最优尺度和非监督评价法的分割及分类结果进行比较,结果表明：该方法能够获取与地面目标相匹配的分割尺度,改善了分割效果,提高了分类精度,具有一定实用价值。     

Rapid growth of angle-confined large-domain graphene bicrystals

In the chemical vapor deposition growth of large-area graphene polycrystalline thin films,the coalescence of randomly oriented graphene domains results in a high density of uncertain grain boundaries (GBs).The structures and properties of various GBs are highly dependent on the misorientation angles between the graphene domains,which can significantly affect the performance of the graphene films and impede their industrial applications.Graphene bicrystals with a specific type of GB can be synthesized via the controllable growth of graphene domains with a predefined lattice orientation.Although the bicrystal has been widely investigated for traditional bulk materials,no successful synthesis strategy has been presented for growing two-dimensional graphene bicrystals.In this study, we demonstrate a simple approach for growing well-aligned large-domain graphene bicrystals with a confined tilt angle of 30° on a facilely recrystallized single-crystal Cu (100) substrate.Control of the density of the GBs with a misorientation angle of 30° was realized via the controllable rapid growth of subcentimeter graphene domains with the assistance of a cooperative catalytic surface-passivation treatment.The large-area production of graphene bicrystals consisting of the sole specific GBs with a tunable density provides a new material platform for fundamental studies and practical applications.     

A novel de-rusting method with molten salt precleaning and laser cleaning for the recycling of steel parts

Clean Technologies and Environmental Policy - Steel parts are a kind of high recovery value product for which de-rusting is the primary step in the recycling process. Laser cleaning is a popular...