国内生存分析在可靠性数据分析中的应用进展

针对可靠性数据分析中遇到的小样本、异总体、删失和截尾问题,引入现代统计学的重要分支——生存分析理论,分别从参数生存分析模型、非参数生存分析模型和半参数生存分析模型3个方面分析了国内生存分析在可靠性数据分析中的研究现状与应用进展。     

小型堆不可隔离小破口事故下的舱室放射性活度可视化

以某小型压水堆为研究对象,运用MELCOR程序建立事故分析模型,研究了不可隔离小破口合并全船断电始发严重事故下舱室放射性活度变化规律。结果表明:不可隔离小破口事故下,倘若安全设施失效,将导致堆芯熔化;堆舱放射性活度比其相邻舱室高4个量级。在此基础上,开发了小型堆典型严重事故下辐射场可视化系统,可为舱室剂量评估和核应急决策提供支持。     

“What is dear to you?” Survey of beliefs regarding protection of critical infrastructure against terrorism

Up to the present, there is only very little research on how the population perceives terrorism and its threats, even though support from the population is crucial for effective counterterrorism. By eliciting beliefs and subjecting them to content analyses, six factors were found that determine the protection worthiness of a target in the people's view: the potential damage to “people,” “symbolism,” “economy,” “politics,” “nature,” and “image/publicity.” These empirically found factors are in line with factors specified by terrorist target selection models. They differ in the strength of their cognitive representation among participants and, thus, their subjective importance to the people. The first three factors are shared among all participants, whereas the latter ones could only be found in a part of the participant sample. People's judgments of the targets' protection worthiness differ substantially from their judgments of the targets' attractiveness to terrorists, even though the same factors seem to be involved. This study offers an insight into the people's mental model about protection worthiness of targets. Together with classical risk analysis and knowledge about terrorists, these results can form a basis for setting up a holistic scheme for critical infrastructure protection.     

融合图嵌入的Smooth-PCANet图像识别算法

主成分分析网络（Principal Component Analysis Network,PCANet）是一种简单的深度学习算法,在图像识别领域具有优秀的性能。然而PCANet在构建网络卷积核时只关注了图像的主分量信息,忽视了近邻像素点之间的位置关系。而通常情况下,图像的相邻像素点具有空间强相关性,因此利用图结构保持像素点的位置近邻关系更有利于网络提取有效特征。因此,我们将图嵌入思想融入PCANet,提出一种新的图像识别算法Smooth-PCANet。为了验证Smooth-PCANet算法的有效性,我们在人脸、手写体字符以及图片等不同数据集上构建实验,并将Smooth-PCANet与多种基于深度学习的图像识别算法作了对比。实验结果证明,Smooth-PCANet算法比PCANet获得了更高的识别性能,并且更有效地避免了过拟合,在小样本训练时具有显著优势。     

Resistance of grid steel-tube-confined concrete targets against projectile impact: Experimental investigation and analytical engineering model

Steel-tube-confined concrete (STCC) targets are provided with excellent anti-penetration performance over semi-infinite concrete (SIC) targets since the steel tube imposes passive restraint on the in-filled concrete during the penetration process. Grid STCC system with square steel tubes is a potential solution to protective structures. In this paper, experiments of 9-cell grid STCC targets penetrated by 12.7 mm Armor Piercing Projectile (APP) were performed. The influence of side length and thickness of steel tube, steel ratio and impact velocity on anti-penetration performance were taken into account. Additionally, single-cell square STCC targets were also designed and tested for comparison with the 9-cell grid STCC targets. Damage modes and parameters of the tested targets were measured and discussed. Moreover, the stiffness of radial confinement of grid STCC targets is achieved according to the elastic solution of infinite cylindrical shell in Winkler medium. Furthermore, the penetration resistance and depth of penetration (DOP) for grid STCC targets are obtained on the basis of the dynamic finite spherical cavity-expansion (FSCE) models including radial confinement effect. It is shown that the 9-cell grid STCC targets with optimal dimension match of thickness and side length of steel tube can reduce the DOP by about 17 % and 23 % in comparison with the SIC targets and single-cell square STCC targets, respectively, due to both the confinement of square steel tube to concrete in the impacted cell and the additional confinement of the surrounding cells to the impacted cell; the penetration resistance and DOP of the grid and cellular STCC targets with similar steel ratio is close, and thus the grid STCC targets with simpler manufacturing process and excellent in-plane expandability are preferred in engineering practice; moreover, the predicted results of DOP model based on the FSCE models agree well with the tested results with the maximum disparity less than 12 % and the proposed model is more applicable to the grid and cellular STCC targets with high radial confinement.     

A pixel-level local contrast measure for infrared small target detection

Infrared (IR) small target detection is one of the key technologies of infrared search and track (IRST) systems. Existing methods have some limitations in detection performance, especially when the target size is irregular or the background is complex. In this paper, we propose a pixel-level local contrast measure (PLLCM), which can subdivide small targets and backgrounds at pixel level simultaneously. With pixel-level segmentation, the difference between the target and the background becomes more obvious, which helps to improve the detection performance. First, we design a multiscale sliding window to quickly extract candidate target pixels. Then, a local window based on random walker (RW) is designed for pixel-level target segmentation. After that, PLLCM incorporating probability weights and scale constraints is proposed to accurately measure local contrast and suppress various types of background interference. Finally, an adaptive threshold operation is applied to separate the target from the PLLCM enhanced map. Experimental results show that the proposed method has a higher detection rate and a lower false alarm rate than the baseline algorithms, while achieving a high speed.     

一种有效的红外弱小目标DBT算法研究

作为红外自寻的制导、搜索跟踪和预警等领域的一项关键技术,红外弱小目标检测与跟踪成了红外图像处理领域中的一项重要研究课题。本文采用了"先检测后跟踪(DBT)"的思想对红外小目标进行检测与跟踪,首先采用基于各向异性偏微分方程的背景抑制技术对单帧图像进行抑制,再用最大绝对对比度阈值对红外图像进行分割,最后采用形心跟踪法对小目标点坐标定位。试验证明本文算法应用于目标检测的可行性和有效性。     

反海盗任务中大中口径舰炮火力应用

在反海盗任务中,大中口径舰炮武器系统可担负对海盗船进行警示或警告性射击,或者直接打击的任务。在对海盗船编队目标特性分析的基础上,以单管大口径舰炮对集群海盗船射击作为研究对象,将射击类型分为警示、警告和阻拦射击,重点分析了对集群海盗船只警告性射击方式,提出了警告性射击修正量的计算方式,最后通过实例仿真证明了该方法的适用性。     

一种新的避免航迹合并的比例联合概率关联算法

通过对传统的联合概率数据关联算法(JPDA)在平行临近及小角度交叉目标的关联跟踪中引起的航迹合并问题的分析,提出了一种基于最大交互概率的改进比例联合概率数据关联算法.该算法找出距离两逼近目标的交互概率中心最近的有效观测,引入一比例因子对该观测与所有目标的关联概率进行修正,将逼近目标“拉开”,从而抑制了航迹的合并.仿真结...     

基于间歇采样转发干扰的ISAR群目标生成方法

基于间歇采样转发干扰理论,提出了一种ISAR二维群目标生成方法。采用伴飞式干扰机,对所掩护的目标回波信号进行间歇采样转发,在ISAR上形成逼真的虚假二维群目标图像;通过调整干扰机和目标的空间位置关系,生成二维假目标序列,并能模拟目标的动态姿态变化。在介绍了间歇采样转发干扰的基本原理后,阐述了伴飞式干扰原理,然后从理论上推导了了伴飞式干扰机的功率和空间位置要求,接下来着重分析了二维群目标生成的理论原因,指出了影响二维群目标分布的关键因素和群目标的姿态角变化规律,最后进行了仿真验证,结果表明了该方法的有效性。研究成果对于ISAR干扰机的设计和工程应用具有指导意义。