可见光与红外图像融合质量评价指标分析

目的 客观评价作为图像融合的重要研究领域,是评价融合算法性能的有力工具。目前,已有几十种不同类型的评价指标,但各应用领域包括可见光与红外图像融合,仍缺少统一的选择依据。为了方便比较不同融合算法性能,提出一种客观评价指标的通用分析方法并应用于可见光与红外图像融合。方法 将可见光与红外图像基准数据集中的客观评价指标分为两类,分别是基于融合图像的评价指标与基于源图像和融合图像的评价指标。采用Kendall相关系数分析融合指标间的相关性,聚类得到指标分组;采用Borda计数排序法统计算法的综合排序,分析单一指标排序和综合排序的相关性,得到一致性较高的指标集合;采用离散系数分析指标均值随不同算法的波动程度,选择充分体现不同算法间差异的指标;综合相关性分析、一致性分析及离散系数分析,总结具有代表性的建议指标集合。结果 在13对彩色可见光与红外和8对灰度可见光与红外两组图像源中,分别统计分析不同图像融合算法的客观评价数据,得到可见光与红外图像融合的建议指标集(标准差、边缘保持度),作为融合算法性能评估的重要参考。相较于现有方法,实验覆盖20种融合算法和13种客观评价指标,并且不依赖主观评价结果。结论...     

小目标检测研究综述

随着计算机视觉和人工智能技术的快速发展,目标检测受到了更加广泛的关注。由于小目标像素占比小、语义信息少、易受复杂场景干扰以及易聚集遮挡等问题,导致小目标检测一直是目标检测领域中的一大难点。目前,视觉的小目标检测在生活的各个领域中日益重要。为了进一步促进小目标检测的发展,提高小目标检测的精度与速度,优化其算法模型,本文针对小目标检测中存在的问题,梳理了国内外研究现状及成果。首先,分别从小目标可视化特征、目标分布情况和检测环境等角度对小目标检测的难点进行了分析,同时从数据增强、超分辨率、多尺度特征融合、上下文语义信息、锚框机制、注意力机制以及特定的检测场景等方面系统总结了小目标检测方法,并整理了在框架结构、损失函数、预测和匹配机制等方面发展的较为成熟的单阶段小目标检测方法。其次,本文对小目标检测的评价指标以及可用于小目标检测的各类数据集进行了详细介绍,并针对部分经典的小目标检测方法在MSCOCO(Microsoft common objects in context)、VisDrone2021(vision meets drones2021)和Tsinghua-Tencent100K等数据...     

Circularly polarized patch antenna with simultaneously wide frequency tuning range and high gain performance

A tunable circularly polarized square patch antenna with parasitic elements is designed for a wide frequency tuning range and high gain characteristics. The proposed antenna is constructed by one main patch and four semi-elliptic parasitic units. By loading four varactor diodes and adjusting their capacitance values, the tunable feature is performed to reallocate the corresponding working frequency. Moreover, the diagonal corners of the antenna are cut and loaded with varactor diodes, which provide the appropriate perturbation between the two orthogonality modes of the antenna, so as to ensure the circular polarization characteristic in the entire operating tuning band. The experimental results demonstrate that the reflection coefficient and axial ratio are less than −13 dB and 3 dB, respectively. The proposed antenna features a relatively wide continuously tuning range of 24% within 1.9-2.3 GHz and a stable gain of over 7 dBi with a radiation efficiency of above 85%.     

基于DeblurGAN的运动模糊图像盲复原算法研究

DeblurGAN方法利用条件生成对抗网络解决了端到端的图像去模糊问题,但存在图像边缘细节恢复不足以及鲁棒性不高的问题,针对此问题,提出一种基于DeblurGAN的运动模糊图像盲复原方法。在生成网络中,采用多尺度卷积核神经网络提取特征,并使用级联空洞卷积扩大神经元的感受野;采用自适配归一化方法代替原来生成器中使用的实例归一化方法。其次,引入了梯度图像L1损失,结合对抗损失和感知损失,将其作为图像去模糊的正则约束,使得生成图像的边缘特征更加清晰。实验结果表明,提出方法复原的图像峰值信噪比数值较DeblurGAN算法高出5.4%,结构相似性指标高出1%;在主观上清晰化效果较好,且消除了网格效应。     

自适应加权的多特征融合ECT图像重建算法

为了解决电容层析成像(electrical cpacitance tomography, ECT)图像重建中电容值与介电常数这一非线性病态问题,提出一种自适应加权的多特征融合(adaptive weighted multi-feature fusion, AWMF) ECT图像重建算法,利用网络模型拟合电容张量和介电常数的非线性映射关系。首先,在网络模型中采用密集卷积网络(DenseNet),不仅缓解梯度消失现象,还融合不同通道的特征信息;添加挤压激励网络(squeeze excitation network, SENet)自适应调整特征通道的权重,用以提取不同通道的关键特征,提高重建图像的精度。其次,构建树形聚合结构（tree aggregation structure network, TASN）网络模块,扩大感受野并提取丰富的多尺度特征信息,消除普通卷积所带来的伪影现象。在COMSOL5.3软件上建模仿真后,通过MATLAB2014a对图像进行重建。实验结果表明,重建图像误差系数降低至0.0256,相关系数提高至0.9717,与传统算法和CNN算法相比,具有更高的图像重建质量。     

On measuring network robustness for weighted networks

Knowledge and Information Systems - Network robustness measures how well network structure is strong and healthy when it is under attack, such as vertices joining and leaving. It has been widely...     

Saliency guided self-attention network for pedestrian attribute recognition in surveillance scenarios

Pedestrian attribute recognition is often considered as a multi-label image classification task. In order to make full use of attribute-related location information, a saliency guided sel-attention network ( SGSA-Net) was proposed to weakly supervise attribute localization, without annotations of attribute-related regions. Saliency priors were integrated into the spatial attention module ( SAM ). Meanwhile,channel-wise attention and spatial attention were introduced into the network. Moreover, a weighted binary cross-entropy loss ( WCEL) function was employed to handle the imbalance of training data. Extensive experiments on richly annotated pedestrian ( RAP) and pedestrian attribute ( PETA) datasets demonstrated that SGSA-Net outperformed other state-of-the-art methods.     

双脉冲激光诱导铝等离子体的双波长干涉诊断

双脉冲激光诱导等离子体在激光加工、元素检测、材料去除等领域有广阔的应用前景和发展空间,对其进行诊断具有重要意义。针对延迟双脉冲激光诱导铝等离子体的作用效果和影响机理,采用双波长干涉法对其时间演化规律展开研究。基于马赫-曾德尔干涉仪搭建了双波长干涉诊断系统,得到了双脉冲激光诱导等离子体干涉图。通过对干涉图的处理和分析,得到了等离子体电子密度随双脉冲激光延迟时间的变化规律。结果表明,随着双脉冲激光延迟时间的增加,第二束脉冲激光对等离子体电子密度的增强效果先加强后减弱。其中,双脉冲激光延迟时间为10 ns时,对等离子体电子密度的增强效果最强,在30 ns时刻,其中心区域平均电子密度可达6.49×1019 cm?3,相较于同等能量单脉冲激光诱导等离子体提升了26%。同时研究了延迟时间对第二束脉冲激光作用机制的影响。研究结果为双脉冲激光诱导等离子体的优化方向提供了参考。     

In Situ Management of Ions Migration to Control Hysteresis Effect for Planar Heterojunction Perovskite Solar Cells

As one of the most promising photovoltaic materials, the efficiency of inorganic–organic hybrid halide perovskite solar cells (PSCs) has reached 25.5% in 2020. However, the stability and hysteresis remain primary challenges before it can become a commercial photovoltaic technology. Therefore, those issues have drawn significant attention for photovoltaic applications. In this work, a study of the PSCs hysteresis improvement is presented based on a combination of first-principles simulations, scanning electron microscopy images, and time-dependent photocurrent measurements. It indicates the hysteresis led by the ion migration and accumulation is mainly localized at the two interfaces: one is between electron transport layer and active layer, and the other is between active layer and hole transport layer. Considering the massive defects at the grain boundaries (GBs), they lower the potential barriers significantly. The defect density at GBs is therefore reduced via the in situ passivation of PbI₂ crystals. The hysteresis index is decreased from 22.43% down to 1.04%, and results in an improvement in efficiency from 17.12% up to 20.10%. Following the understanding of defect-induced hysteresis, an approach to improve the hysteresis is provided, which can be integrated into the fabrication process and widely applied to enhance the performance of PSCs.