Hebbian-based neural networks for bottom-up visual attention and its applications to ship detection in SAR images

This paper proposes a bottom-up attention model based on pulsed Hebbian neural networks. The salience of the visual input can be generated through the networks using a simple normalization process, which can be calculated rapidly. Moreover, visual salience in this model can be represented as binary codes that mimic neuronal pulses in the human brain. Experimental results on psychophysical patterns and eye fixation prediction for natural images prove the effectiveness and efficiency of the model. In an arduous task of detecting ships in synthetic aperture radar (SAR) images, there are large amounts of data to be processed in real time. As a fast and effective technique for saliency detection, the proposed model is applied to ship detection in SAR images and its robustness against speckles is further proved.     

High sensitive and selective formaldehyde sensors based on nanoparticle-assembled ZnO micro-octahedrons synthesized by homogeneous precipitation method

Nanoparticle-assembled ZnO micro-octahedrons were synthesized by a facile homogeneous precipitation method. The ZnO micro-octahedrons are hexagonal wurtzite with high crystallinity. Abundant structure defects were confirmed on ZnO surface by photoluminescence. Gas sensors based on the ZnO micro-octahedrons exhibited high response, selectivity and stability to 1–1000 ppm formaldehyde at 400 °C. Especially, even 1 ppm formaldehyde could be detected with high response (S = 22.7). It is of interest to point out that formaldehyde could be easily distinguished from ethanol or acetaldehyde with a selectivity of about 3. The high formaldehyde response is mainly attributed to the synergistic effect of high contents of electron donor defects (Zn_i and V_O) and highly active oxygen species (O²⁻) on the ZnO surface.     

一种安全快速的虹膜识别方法

虹膜识别技术作为一种生物特征识别手段,具有唯一性、稳定性等优点,从而成为当前模式识别和机器学习领域的一个研究热点.目前的虹膜识别方法多数是采用在图像上提取特征点,并将特征点编码为固定长度的特征数据用于匹配的方式,这种方式的特征提取算法复杂,计算量大,而且使虹膜识别系统易受到攻击.为了提高虹膜识别系统的安全性和识别速度,本文提出一种安全快速的虹膜识别方法,在活体虹膜上进行直接匹配,该算法具有识别速度快,准确率高的特点,经实验验证,识别时间为30.1ms,识别率达到98.86%,与经典算法相比较,算法简单,安全性高,具有放缩,平移、旋转不变性.     

基于STM32的手持式测氧仪设计

针对国内传统手持式测氧仪精度低、功耗高的问题,提出了采用嵌入式微控制器STM32F103RBT6作为控制芯片的手持式测氧仪.该仪器通过控制元器件的开断来降低待机功耗,根据所测浓度人工切换系统放大倍数,通过键盘设置时间和报警值等相关参数.算法上采用中位值平均滤波算法进行软件滤波,结合传统两点标定法和多点标定法的优缺点,采用多点标定法进行氧浓度标定,采用分段斜率拟合算法计算氧浓度值.经实验测试证明设计的手持测氧仪具有高精度和高稳定性.     

Robust finite-time boundedness of multi-agent systems subject to parametric uncertainties and disturbances

In this paper, we consider finite-time control problems for linear multi-agent systems subject to exogenous constant disturbances and impulses. Some sufficient conditions are obtained to ensure the finite-time boundedness of the multi-agent systems, which could be then reduced to a feasibility problem involving linear matrix inequalities. Numerical examples are given to illustrate the results.     

爆炸冲击条件下的加速度传感器结构分析

针对加速度传感器在爆炸与冲击测试中的应用,从理论与有限元仿真出发,分析传感器结构的静态响应与冲击响应.在15.4×104gn的静态载荷下,传感器结构最大应力超过材料的许用应力,将会发生结构断裂.在静态载荷下,加速度传感器在15.4×104gn的冲击加速度载荷下结构最大应力超过材料的许用应力,将会发生结构断裂.在加速度传感器的工作方向上施加幅值为15×104gn,半周期为5μs、10μs、20μs、30μs、40μs的半正弦加速度冲击载荷.在幅值为15×104gn、半周期为30μs的冲击载荷下,传感器的固定端处应力为334MPa,将会使传感器断裂失效.在幅值为15×104gn、半周期为5μs、10μs、20μs的冲击载荷下,固定端处应力超过材料许用应力,将也会发生结构断裂.悬臂梁在半周期为5μs、10μs、20μs的冲击下,将会出现断裂.大体上,冲击载荷的周期越小,固定端的应力越大集中越严重.由于传感器固有周期为9.5μs,加速度传感器在半周期为10μs的冲击载荷下出现谐振,固定端处应力变大集中加剧.分析加速度传感器在冲击载荷下的结构响应为传感器的结构设计与具体应用时的可靠性分析提供了理论依据.     

大岗山水电站工程边坡稳定性分析

采用SLOPE/W模块对大岗山水电站工程边坡稳定性进行了计算分析。该水电站右岸工程边坡高达420m,卸荷裂隙控制右岸边坡的整体稳定性,构成确定性的2个潜在不稳定块体。分析结果表明,天然状态下右岸边坡整体稳定,地震对边坡稳定性影响较大,开挖对边坡稳定性有一定影响。     

基于自动调焦的微机控制显微硬度计

传统显微硬度计自动化程序不高,难以满足现代信息处理与分析要求,本文提出的基于自动调焦的微机控制技术应用传统显微硬度计中,能对各种金相试样进行自动测试,实践中取得了很好的效果。     

数字PID调节器应用中的几个技术问题

本文就数字PID调节器应用中的正、反作用、积分饱和、限位及手动/自动无扰动切换等几个技术问题进行了讨论,并提出了具体软件解决办法.     

我国GNP的单位根不确定性研究

对于一个实际的时间序列数据,我们并不知道其真正的数据生成过程,只能通过假设和基于假设的统计推断来确定。本文利用蒙特卡罗模拟进行有限样本的ADF检验。结果表明,当原假设是存在单位根,备择假设是趋势稳定时,ADF检验的势很低。这意味着,我国GNP存在单位根是不确定的。