提高关联成像重构质量的研究

比较基于压缩感知关联成像(CGI)与伪逆关联成像(PGI)两者之间的成像效果差异,探讨形态学权重自适应对关联成像去除噪声的效果。选择不同的图像,通过MATLAB软件开展仿真实验,对目标图像分别采样64、256、512、1 024、2 048、3 000次,首先通过关联成像、基于压缩感知关联成像与伪逆关联成像三种方法重构图像,再对比压缩感知与伪逆两种方法重构图像的效果,以峰值信噪比(PSNR)、相关系数(CC)为量化指标,将基于压缩感知关联成像与伪逆关联成像在不同采样次数下进行对比分析。同时,通过实验分析形态学权重自适应去除关联成像中噪声的效果。伪逆关联成像在低次数采样的情况下比基于压缩感知关联成像的成像效果更好,在高采样次数下,基于压缩感知关联成像的成像效果更好。在实际重构中压缩感知关联成像重构的图像仍有噪声,形态学权重自适应可以有效去除关联成像实验中产生的噪声。     

A Colorimetric Artificial Olfactory System for Airborne Improvised Explosive Identification

The detection of ultralow or nonvolatile target analytes remains a significant challenge for artificial olfactory systems even after decades of development, which severely limits their widespread application. To overcome this challenge, an artificial olfactory system based on a colorimetric hydrogel array is constructed for the first time as a universal representative. As an effective extension of conventional artificial olfactory systems that integrates the merits of its predecessors, the proposed system accurately mimics olfactory mucosa and specific odorant binding proteins using hydrogels endowed with specific colorimetric reagents for the detection of hypochlorite, chlorate, perchlorate, urea, and nitrate. Therefore, the proposed system is capable of detecting and discriminating between these five airborne improvised explosive microparticulates with a detection limit as low as 39.4 pg. Additionally, the system demonstrates good reusability over ten cycles, rapid response time of ≈0.2 s, and excellent discrimination properties, despite significant variation. This proof-of-concept study on colorimetric artificial olfactory systems yields a novel strategy for the direct and discriminative detection of nonvolatile airborne microparticulates.     

BRISOL钠同位素放射性核束的产生

北京放射性核束装置在线同位素分离器（BRISOL）采用100 MeV、200 μA回旋加速器提供的质子束打靶产生中、短寿命放射性核束，在线分析后供物理用户使用，其质量分辨率好于20 000。为开展²⁰Na核的奇异衰变特性研究，研制了氧化镁靶，并采用100 MeV质子束轰击氧化镁靶在线产生了^20～26Na⁺的钠同位素放射性核束。当质子束流强为8 μA时，²⁰Na⁺离子束的最大产额为2×10⁵ s^-1，²¹Na⁺离子束的最大产额为4×10⁸ s^-1。完成了北京放射性核束装置首个放射性核束物理实验，累计供束近200 h。     

氮气对复采工作面遗煤复燃的防治效能

为探究不同体积分数N₂对复采工作面遗煤复燃过程的防治效果，采用自主研发的程序升温试验装置，对煤样分别通入体积分数为10%、20%、30%和40%的N₂，以氧化反应过程中耗氧速率V_O2、一氧化碳产生率V_CO和表观活化能E的变化情况来反映煤自燃的发展程度。试验表明:不同体积分数的N₂对煤氧化升温过程均具有抑制作用，相比于煤在纯空气条件下的氧化自燃，通入的N₂体积分数越大，煤初次自燃和二次氧化升温过程的V_O2和V_CO更低、E更高，说明通入的氮气体积分数越大，对煤氧化升温过程的抑制效率越好。其中，在煤初次自燃阶段，当N2体积分数C_N2≥20%时抑制效果更佳；在煤二次氧化复燃阶段，当N₂体积分数C_N2≥40%时抑制效果更佳。     

综放工作面风量及采空区漏风定量检测

通过对中煤塔山煤矿30509综放工作面采空区漏风及有害气体分析的研究，总结了通过定量连续释放SF₆示踪气体，来检测工作面风量与采空区漏风的方法，测定了沿工作面采空区的漏风分布及漏风流线的深度。采用定时测量法，确定了30509工作面随工作面不断推进，与对应地表发生了导通现象；采用定量测量法，确定了内部漏风量和外部漏风量较大，对综采工作面采空区自然发火存在较大风险，同时提出了防范措施，为减少综放工作面采空区漏风、防止自然发火的发生提供了参考。     

An energy‐dissipating approach for adjacent retrofitted structures

The approach of positioning energy dissipaters between adjacent structures can effectively prevent the collision of two adjacent buildings. For adjacent retrofitted structures, an effective strengthening approach that involves implementing energy‐dissipation measures both between and within buildings is proposed. This paper is based on the actual Shanghai ShiMao International Plaza. To avoid having to strengthen most of the structural members, the strengthening measure of energy dissipation is selected because of its construction cost and time constraints. This approach involves replacing several dampers between the adjacent structures and adding viscous damping walls in the retrofitted structure to reduce the seismic response of the retrofitted structure back to the original level. Additionally, a practical calculating method for the equivalent additional damping ratio for the adjacent structures connected by energy dissipaters is proposed. The results show that the effect of energy dissipation in the retrofitted structure is prominent and that the goal of indirect strengthening is achieved. The strengthening approach of energy dissipation and the calculating method of the equivalent additional damping ratio for adjacent structures provide a reference for similar engineering designs.     

雷达目标辨识网络的对抗性样本分析

为了验证雷达目标辨识网络存在风险,提升基于辨识网络的雷达目标辨识效果,文中研究了不同雷达目标辨识网络的对抗性样本。针对雷达目标辨识网络中的卷积神经网络和分解卷积神经网络,建立对抗性样本生成模型,按照该模型生成对抗性样本,并对生成的结果进行分析总结。实测数据处理结果表明,雷达目标辨识网络存在潜在风险。     

因子分析在冀东峪耳崖金矿地球化学找矿中的应用

峪耳崖金矿位于华北板块北缘，矿区基底地层为太古界迁西群麻粒岩相和角闪岩相，矿体受NNE—NE向断裂构造控制，赋矿围岩为花岗岩。在矿床地质特征基础上，利用矿区土壤地球化学数据，采取传统统计法和累计频率法圈定了金矿12个元素的异常范围；运用因子分析方法划分出了Sb Hg Pb As元素组合、Au Ag Bi元素组合、Zn Cu Sn元素组合和W Mo元素组合；再对因子得分进行趋势面分析，较为有效地突出了深层次的找矿信息，更准确地圈出成矿靶区。结合研究区地质特征认为Au Ag Bi元素组合和Sb Hg Pb As元素组合为研究区最佳地球化学标志元素组合，下一步地质工作重点应放在矿区中矿带矿体深部及其南部异常高值区，同时布置少量工程在北矿带，以追索与唐杖子金矿之间的联系。     

机动小目标的时空联合检测技术

主动声呐信号处理中一般采用匹配滤波和背景均衡等算法检测回波信号,为提高信号的检测性能,还需利用信号与混响和环境噪声在时间、空间上的统计特征差异。文章在对信号的时、空特征算法进行简单描述的基础上,提出了频率能量统计特征和方位能量统计特征,将两种特征检测器的输出进行融合处理构成时、空联合检测器。仿真和试验数据处理结果证明,此方法可有效提高机动小目标在混响及环境噪声中的检测性能。