ASW-RTS:一种软件性能数据在线检测方法

在软件健康管理系统中,需要对相关被测构件的性能进行在线检测,从而确定构件的健康状态。针对被测构件的性能衰退检测问题,在AR(p)模型基础上,结合时间序列滚动式算法的准确性,提出了一种具有自适性策略的ASWRTS算法,用于对被测构件革新异常类性能数据的在线检测,提高检测的准确度,并基于该算法设计了构件性能异常数据的在线检测过程。实验结果表明,该算法将检测精度提高了50%左右,能够有效在线检测构件的性能衰退,说明了该算法的正确性和有效性。     

某火箭发动机故障检测及诊断算法设计分析

针对某型号液体火箭发动机故障检测及诊断软件存在瞬态段（起动段、关机段）故障检测功能缺失、故障诊断无法给出故障量大小的问题，在原软件故障诊断算法——余弦相似度分类的基础上，通过引入故障因子定量给出故障量大小，实现发动机故障诊断的定性定量分析。同时，采用基于统计学基础的包络线算法，通过Python语言对发动机瞬态段故障检测算法进行设计开发，使原软件实现发动机瞬态段故障检测功能。通过仿真试车数据对软件算法进行验证，结果表明优化后的故障诊断功能可实现故障量大小计算，基于包络线算法开发的故障检测功能提高了发动机故障检测效率。     

Hypothesis-test-based landcover change detection using multi-temporal satellite images – A comparative study

Remote sensing images and technologies have been widely applied to environmental monitoring, in particular landuse/landcover classification and change detection. However, the uncertainties involved in such applications have not been fully addressed. In this paper two hypothesis-test-based change detection methods, namely the bivariate joint distribution method and the conditional distribution method, are proposed to tackle the uncertainties in change detection by making decisions based on the desired level of significance. Both methods require a data set of class-dependent no-change pixels to form the basis for class-dependent hypothesis test. Using an exemplar study area in central Taiwan, performance of the proposed methods are shown to be significantly superior to two other commonly applied methods (the post-classification comparison and the image differencing methods) in terms of the overall change detection accuracies. The conditional distribution method takes into consideration the correlation between digital numbers of the pre- and post-images and the effect of the known pre-image digital number on the range of the post-image digital number, and therefore yields the highest change detection accuracy. It is also demonstrated that the class-dependent change detection is crucial for accurate landuse/landcover change detection.     

A salient edges detection algorithm of multi-sensor images and its rapid calculation based on PFCM kernel clustering

Multi-sensor image matching based on salient edges has broad prospect in applications, but it is difficult to extract salient edges of real multi-sensor images with noises fast and accurately by using common algorithms. According to the analysis of the features of salient edges, a novel salient edges detection algorithm and its rapid calculation are proposed based on possibility fuzzy C-means （PFCM） kernel clustering using two-dimensional vectors composed of the values of gray and texture. PFCM clustering can overcome the shortcomings that fuzzy C-means （FCM） cluster- ing is sensitive to noises and possibility C-means （PCM） clustering tends to find identical clusters. On this basis, a method is proposed to improve real-time performance by compressing data sets based on the idea of data reduction in the field of mathematical analysis. In addition, the idea that kernel-space is linearly separable is used to enhance robustness further. Experimental results show that this method extracts salient edges for real multi-sensor images with noises more accurately than the algorithm based on force fields and the FCM algorithm; and the proposed method is on average about 56 times faster than the PFCM algorithm in real time and has better robustness.     

基于模块度的复杂网络社团结构聚类新算法

针对传统的分裂聚类算法删除连边会影响节点划分的问题,结合K-means算法的思想并加以改进,提出了一种以邻居节点为聚类备选集,所有社团中心节点同时聚类的同步聚类算法。该算法根据节点中心度和最短路变化率确定社团中心节点集,然后以中心节点为社团中心,以邻居节点为聚类备选集合进行聚类,完成社团的划分。将算法应用到Zachary网络中并与GN算法、Top Leader算法进行比较,仿真结果表明该算法对网络有较好的划分。     

单天基平台测向的空间非合作目标跟踪

针对仅含角度测量信息的单个天基平台可观测性较弱的问题,提出了一种含脉冲机动检测的空间非合作目标跟踪算法,并设计了非合作目标实时跟踪数据处理流程.该算法利用抗差估计技术和UKF（Unscented Kalman Filter,无迹卡尔曼滤波）算法构造目标跟踪滤波器,并综合残差多项式拟合和新息分布特征等方法实现目标机动检测,在天基平台观测信息类型有限和观测几何较差的情况下,可以同时排除孤立野值和成片测量野值的影响,实现非合作机动目标的连续稳定跟踪.数值实验验证了算法的可行性和有效性,也表明了跟踪精度和可靠性与测量精度密切相关.     

基于概率统计的DS/FH系统干扰检测算法

基于时频分布的干扰检测算法是DS/FH（Direct Sequence/Frequency Hopping,直接序列扩频/跳频扩频）系统干扰检测的传统手段,针对其运算量大、实时性差,以及无法检测某些特殊干扰信号等问题,提出一种基于概率统计的干扰检测算法.通过统计DS/FH系统频谱图中各频率的幅值出现位置,以DS/FH信号与干扰信号频谱分布规律的差别为判决依据,完成DS/FH系统中干扰信号的检测.分析和仿真结果表明：该检测算法能够有效识别DS/FH系统中的典型干扰信号,与传统基于时频分布的检测算法相比,能够实现某些特殊干扰信号的有效检测;同时,由于该算法不需要进行复杂的时频联合分析,其运算复杂度大大降低,能够实现DS/FH系统干扰信号的快速检测.     

海战场电子干扰环境下雷达作用距离建模与仿真

针对如何模拟海战场电子干扰环境下雷达作用距离这一问题,首先,分析了雷达检测因子与虚警概率、检测概率的关系;然后,在考虑电子干扰功率的基础上建立了综合信干比模型,利用综合信干比与检测因子之间的关系确定了雷达作用距离;最后,通过仿真分析,验证了模型的有效性和适用性。     

面向天基监视的红外弱小飞行目标识别算法

针对当前红外弱小飞行目标特征不明显、背景干扰大等问题,提出了一种基于深度学习的红外弱小目标识别算法。检测框架以YOLOv4模型为基础,通过使用K-means++算法对训练集的候选框进行聚类处理,在初始大小的选取上放弃随机生成初始点的方式,在样本集里选取某一个样本作为初始中心使锚框（anchor）大小的选取更加合理。在模型结构中引入卷积注意力模块,使算法模型计算资源分配更合理,对红外弱小飞行目标的特征信息更加敏感。改进空间金字塔池化模块,使用平均池化可以更多保留图像的原始信息,降低天基成像中的噪点与坏点的影响。仿真实验表明采用K-means++计算Anchor大小时准确率可以达到80.13%,在加入了SPP和CBAM模块后之后在测试集上算法识别准确率达到了83.3%,经过对模型的修改有效提升了对红外弱小飞行目标识别的准确率。     

基于YOLOv5的近岸SAR舰船目标检测方法

由于类似舰船的陆地目标的干扰和舰船的紧密排列,基于合成孔径雷达(SAR)图像的近岸舰船检测会出现较多漏检和误检,提出了一种基于YOLOv5网络的近岸舰船的检测方法。为提高近岸背景下的检测精度,使用了注意力机制模型和CSL的技术用于改进网络;分析了YOLOv5网络、注意力模型和CSL算法,基于YOLOv5进行了检测实验,引入注意力模型来改进网络;结合CSL算法,重构了YOLOv5旋转检测网络。通过调整训练参数和改进注意力,近岸目标检测网络的测试结果达到mAP 80%以上,证实了CSL+YOLOv5算法实现旋转检测的可行性。