微分Haar-Gaussian小波变换在大景深物体边缘测量中的应用

提出了微分Haar_Gaussian小波变换算法并将其应用于边缘检测,在空域和频域中进行变换,并应用带宽匹配方法和远心测量光学系统,提高了运算速度,通过对实际大景深物体离焦像的边缘测量,得到了准确的结果。     

基于Radon变换的横向剪切干涉图的自动检测技术

在分析测试原理的基础上,指出了现有方法的横向剪切量获得的不足,根据剪切图和Radon变换在边缘处具有突变的特点,提出了一种基于Radon变换的横向剪切干涉图的剪切量的自动检测方法,首先根据二维图像的Radon变换结果,得到剪切图的横向移动方向,对图形进行方向校正。然后利用干涉图形的特点对图形在X和Y方向上进行Radon变换,得到剪切量。利用MATLAB进行编程,实验结果表明该方法具有很好的性能和较高的实用价值。     

基于多DSP的车轮外形磨耗动态检测系统

针对矿井绞车车轮外形磨耗动态检测问题,构建了由线状光源、CCD、多DSP图像处理子系统等组成的系统.对于车轮磨耗图像的获得和外形曲线的提取,重点阐述了多DSP处理系统在绞车车轮外形磨耗实时检测中的设计问题.车轮外形的提取采用了一种梯度边缘检测算法来实现,并给出了磨耗值计算的变换矩阵.初步试验表明该系统的检测误差不大于0.3 mm.     

机车轮对轴颈高精度动态检测技术的研究

通过对线阵CCD成像系统的研究,提出了机车轮对轴颈磨损的CCD成像拼接动态检测方法,分析了该方法的可行性.讨论了斜光束照射、背景光强变化等因素,这些因素降低了系统的测量精度.最后通过实验系统证明了该方法的可靠性.     

基于贝叶斯网络的船舶港口国监督检查滞留风险研究

针对港口国监督(Port State Control, PSC)检查的复杂性和不确定性, 基于贝叶斯网络理论构建船舶PSC检查滞留风险分析模型. 以东京备忘录(Tokyo MOU)中2014~2017年船舶PSC检查样本数据为基础, 运用R语言bnlearn包进行贝叶斯网络的结构及参数学习. 同时分别执行贝叶斯网络的正向、逆向推理, 定量表示各风险因素与滞留结果之间的相互作用关系, 找出导致船舶滞留的高风险因素, 实现不确定环境下船舶PSC检查滞留风险的全面动态分析. 实证表明, 模型具有较高的精确度, 可为检查人员的滞留决策及航运公司的安全风险管理提供有效依据.     

Rectifying resistance switching behaviors of SnO2 microsphere films modulated by top electrodes

Based on the bipolar resistive switching (RS) characteristics of SnO₂ films, we have fabricated a new prototypical device with sandwiched structure of Metal/SnO₂/fluorine-doped tin oxide (FTO). The SnO₂ microspheres film was grown on FTO glass by template-free hydrothermal synthesis, which was evaporated with various commonly used electrodes such as aluminium (Al), silver (Ag), and gold (Au), respectively. Typical self-rectifying resistance switching behaviors were observed for the RS devices with Al and Au electrodes. However, no obvious rectifying resistance switching behavior was observed for the RS device with Ag electrode. Above results were interpreted by considering the different interface barriers between SnO₂ and top metal electrodes. Our current studies pave the ways for modulating the self-rectifying resistance switching properties of resistive memory devices by choosing suitable metal electrodes.     

Mathematical models for job-shop scheduling problems with routing and process plan flexibility

As a result of rapid developments in production technologies in recent years, flexible job-shop scheduling problems have become increasingly significant. This paper deals with two NP-hard optimization problems: flexible job-shop scheduling problems (FJSPs) that encompass routing and sequencing sub-problems, and the FJSPs with process plan flexibility (FJSP-PPFs) that additionally include the process plan selection sub-problem. The study is carried out in two steps. In the first step, a mixed-integer linear programming model (MILP-1) is developed for FJSPs and compared to an alternative model in the literature (Model F) in terms of computational efficiency. In the second step, one other mixed-integer linear programming model, a modification of MILP-1, for the FJSP-PPFs is presented along with its computational results on hypothetically generated test problems.     

计量调整型抽样检验系统GB/T6378-1986与ISO3951 :1989的比较分析

GB/T6378-1986作为计量抽样检验方面的基础性国家标准,是参照国际标准ISO3951:1989而制定的。1989年国际标准化组织ISO/TC69发布了ISO3951的修订版本ISO3951:1989,ISO3951的这一修订版反映了计量抽样检验的最新研究成果,为开展国家标准GB/T6378的修订工作,本文系统地对GB/T6378-86和ISO3951:89进行了比较分析研究。     

自动实时激光全息无损检测系统

介绍一种自动激光全息无损检测系统。该系统由计算机控制自动曝光，并对全息干版进行自动冲洗和对测量数据进行处理，解决了一次曝光激光全息实时检测中全息干版无法精确复位的问题。计算机控制真空加载，缺陷大小及位置由计算机打印输出，一次最大测量面积为０５ｍ×１２ｍ，最小分辨缺陷尺寸为３ｍｍ。     

Technically optimal inspection policy with arithmetical adaption

A process is considered whose quality deteriorates accordingto a constant failure intensity . As in practice it can be difficultto estimate the true value of the purpose of this paper isto present a strategy which can be applied without knowing .In order to maximize the number of conforming items per timeunit perfect inspections and renewals are performed. The lengthof the inspection interval is described by an arithmetical sequenceand changes by the time depending on perceived assignable causes.Optimal adaptive control plans provide nearly the same performanceas in the case when is known.