融合神经网络与超像素的候选区域优化算法

为解决目标检测中候选区域召回率低的问题,提出融合神经网络与超像素的目标候选区域算法。该算法利用神经网络提取更能清楚表达目标边界的特征,并使用聚类、相似性等策略,计算每个滑动窗口所含有的边缘信息量;将待测图像使用简单线性迭代聚类算法分割成若干个超像素,并利用超像素的空间位置、完整性、相邻超像素间的对比度信息,计算各个超像素的显著性得分及每个滑动窗口的显著性得分;根据每个滑动窗口的边缘信息及显著性得分筛选滑动窗口。在PASCAL VOC 2007测试集上进行对比实验,其实验结果表明:所述算法能够快速产生定位质量高的候选区域。     

Study on energy release characteristics of reactive material casings under explosive loading

Reactive Materials (RMs), a new material with structural and energy release characteristics under shock-induced chemical reactions, are promising in extensive applications in national defense and military fields. They can increase the lethality of warheads due to their dual functionality. This paper focuses on the energy release characteristics of RM casings prepared by alloy melting and casting process under explosive loading. Explosion experiments of RM and conventional 2A12 aluminum alloy casings were conducted in free field to capture the explosive fireballs, temperature distribution, peak overpressure of the air shock wave and the fracture morphology of fragments of reactive material (RM) warhead casings by using high-speed camera, infrared thermal imager temperature and peak overpressure testing and scanning electron microscope. Results showed that an increase of both the fireball temperature and air shock wave were observed in all RM casings compared to conventional 2A12 aluminum ally casings. The RM casings can improve the peak overpressure of the air shock wave under explosion loading, though the results are different with different charge ratios. According to the energy release characteristics of the RM, increasing the thickness of RM casings will increase the peak overpressure of the near-field air shock wave, while reducing the thickness will increase the peak overpressure of the far-field air shock wave.     

航天装备大数据资源平台构建技术研究

航天装备研制涉及论证、研制、生产、试验、综合保障等各个环节,所产生的数据具有种类多、数据量大、异构性强等特点,且数据资源分散于各科研机构院所,形成一个个数据孤岛,严重制约了型号数据的统一管理和协同应用。为了建立“逻辑统一、物理分散”的航天装备统一大数据中心,本文提出一种航天装备大数据资源平台构建技术,该平台基于统一的数据编目规范,综合利用云计算、微服务、区块链等技术,可实现多维异构数据深度融合管控,实现基于型号科研生产活动下的数据逻辑统一管理和实体分布式存储,能够解决航天装备整个研制活动全要素全过程的数据采集和可控共享的现实需求,为各业务场景下的数据分析奠定基础。     

一体化联合作战装备保障模式构想

为迎接世界新军事革命的挑战,推进中国特色的军事变革,适应一体化联合作战这一信息时代的作战样式,文章针对一体化联合作战对装备保障的影响与提出的新要求,立足我军现实情况,尝试构建了一体化联合作战装备保障模式,对我军装备保障模式改革进行了一定探索。     

基于DSP的车辆数字电压调节器设计

阐述了一种适用于特种车辆的基于DSP的数字电压调节器的总体设计,从调压器的硬件体系、算法设计和软件体系等方面进行了详细的论述。该调节器采用PID控制技术,充分发挥了DSP高速运算的特点,工作性能稳定可靠,控制精度高。     

基于虚拟样机技术的坦克扭力轴疲劳强度分析

采用虚拟样机技术,应用ADAMS软件中的ATV履带车辆工具箱、实体建模软件(SolidWorks)和有限元分析软件(Patran和Nastran),对柔性化的坦克悬挂系统扭力轴在动载荷工况下的疲劳强度进行了分析,找出了扭力轴的薄弱环节。     

基于时域平均的齿轮故障诊断

提出了信号时域平均处理的新算法,解决了周期截断误差对平均结果的影响,并对测取的齿轮箱齿轮故障振动信号进行了处理。实验结果表明:该方法在齿轮故障信号的处理中能够消除背景噪声的影响,取得比较理想的效果。     

边界扫描测试在显示技术中的应用

为有效检测电路板焊接质量与器件质量,实现高精度的故障定位,提出了基于边界扫描技术与功能测试技术的解决方案。在显示产品设计中,开展可测试性设计,形成完整的边界扫描测试链;针对VRAM器件特点,提出了基于DRAM的测试方法,并设计相应的测试模型。最终,实现了对电路板的边界扫描测试,快速完成了对显示器件质量与焊接质量的有效检测。边界扫描技术与功能测试的结合,可以扩展边界扫描技术的应用范围,实现了更高的测试覆盖率。     

Do Military Expenditures Crowd-out Health Expenditures? Evidence from around the World, 2000–2013

Abstract

The paper examines the relationship between health and military expenditures using pooled cross-sectional (197 countries) and time series (2000–2013) data. Simultaneous equation models were employed to estimate the relationship between an array of public sector expenditures in order to address potential endogeneity. Our empirical findings strongly support the crowding-out hypothesis whereby increased military expenditures reduce the capacity of government to direct expenditures to health expenditures. These findings were robust to alternative specifications explored in the sensitivity analyses. Compared with upper-middle-income countries, the crowding-out effect became more pronounced among lower-middle-income countries. Consequently, this study shows that increased military expenditures negatively impacts health expenditures, and therefore poses as an important risk factor for population health and individual well-being. Moreover, it is the poorest of nations that are most sensitive to the negative effects of increased military expenditures.