试析图书馆网络采访

分析了图书馆传统采访方式的缺点，叙述了网络采访模式的特点和工作流程，指出网络采访是图书馆采访工作的发展方向。     

Development of the CAS-IBB single-particle microbeam for localized irradiation of living cells

A single-particle microbeam facility has been constructed at the Key Laboratory of Ion Beam Bioengineering (LIBB), Chinese Academy of Sciences (CAS). The system was designed to deliver a defined numbers of hydrogen ions, produced by a van de Graaff accelerator, in an energy range of 2.0-3.0 MeV, into an area smaller than that of the nucleus of an individual living cell. The beam is collimated by a borosilicate glass capillary that forms the beam-line exit. An integrated computer program recognizes the cells and locates them one by one over the microbeam exit for irradiation. We present technical details of the CAS-LIBB microbeam facility, particularly on the collimator, hardware, control program, as well as cell irradiation protocols available. Various factors contributing to the targeting and positioning precision are discussed along with accuracy measurement results.     

多传感器机器人数据采集与融合系统研究

根据多传感器机器人控制器数据采集量大、数据融合算法复杂、系统实时性强的特点 ,提出了基于ADμC812和TMS32 0F2 0 6的双CPU结构的机器人信息采集与数据融合系统结构 .介绍了该系统的硬件组成、基本工作原理以及接口电路 ,并给出了ADμC812子系统程序流程图 .     

基于压缩感知的实时在线目标追踪

在正负样本区域随机抽取了不同尺度下图像的局部二值模式(LBP)特征,将高维的特征信息投射到低秩的压缩域,并据此建立了表观模型.使用一个随机稀疏测量矩阵来压缩前景和背景目标.将追踪问题转化成为了一个使用朴素贝叶斯分类器的二元分类问题.所提方法可以较快速、实时地在线追踪目标,同时解决了目标尺度变化、遮挡问题.     

全球气候治理格局下中国低碳技术的发展取向研究

本文主要通过文献研究、政策调研的方法,梳理了中国参与全球气候治理下的低碳技术发展形势。研究发现,中国未来面临的减排压力较大,发展低碳技术势在必行。但是,由于受国际经济利益博弈格局的影响,以及贸易保护主义盛行,实际低碳技术转移效果很差,未来也很难实现突破,所以发展自主研发的低碳技术产业是我国未来的发展方向。中国应积极提升低碳技术专利质量、完善研发模式并实现技术的知识产权战略布局。     

基于SSD卷积网络的视频目标检测研究

针对传统卷积神经网络对远距离视频目标识别效果差的原因,本文提出一种改进的基于SSD卷积网络的视频目标检测模型.首先,对数据集进行剪裁,旋转等预处理,提高网络检测泛化能力,其次,采用coco数据集Mobilenet_SSD预训练模型,由于其具有轻量级网络模型特点,减少计算开销,减少内存占用量.然后,再结合voc2012数据集进行二次训练微调处理,加快训练收敛速度,使用自定义数据集能有效检测特定场景目标,能够有效识别远距离场景下视频目标物体.实验结果表明,改进的网络检测模型适用于远距离目标检测,减少计算量,降低硬件内存资源消耗,提高网络模型性能和检测精确度,具有较好的鲁棒性.     

不同Alpha-Beta滤波算法的精度分析及改进

机动目标跟踪是雷达应用系统的一个重要组成部分,在机动目标跟踪中,目标的跟踪精度作为一项能够反映跟踪算法好坏的指标,是大部分人所关注的重点内容.通过对四种传统的Alpha-Beta滤波算法进行精度分析和比较,总结出四种算法的滤波特点,根据这些特点,改进得出了一种自适应Alpha-Beta滤波算法.该算法相对于传统的Alpha-Beta滤波算法来说,整体滤波效果是最优的,精度也是整体最高的.     

Observation of defect clustering in Gd-doped calcium fluoride

Abstract

Solid solutions Ca_1-xGd_xF_2+x for 3 × 10^?7≤ x ≤10^?1 have been studied by electron paramagnetic resonance (EPR) and ionic thermal currents (ITC). The EPR experiments show the presence of two single-ion sites a cubic and a tetragonal Gd³⁺ center which co-exist with comparable abundances for intermediate impurity concentrations. The cubic center predominates at very low and high concentrations. Seven different relaxation processes have been identified from the ITC spectra and the variation of their intensity vs. x was measured. The absolute concentrations of the cubic and nn Gd³⁺ dipoles were calculated. The scavenging of interstitial fluorines by the neutral clusters explains both the abundance of cubic sites at high concentration and the variety of orientable clusters detected by ITC.     

Characterization of the road surfaces in real time

The main factor in the propagation of traffic noise is the road surface, where the vehicles generate noise due to the contact between tire and pavement, in addition to the noise produced by the engine.     

Practical guide for selection of 1H qNMR acquisition and processing parameters confirmed by automated spectra evaluation

In our recent paper, a new technique for automated spectra integration and quality control of the acquired results in qNMR was developed and validated (Monakhova & Diehl, Magn. Res. Chem. 2017, doi: 10.1002/mrc.4591 ). The present study is focused on the influence of acquisition and postacquisition parameters on the developed automated routine in particular, and on the quantitative NMR (qNMR) results in general, which has not been undertaken previously in a systematic and automated manner. Results are presented for a number of model mixtures and authentic pharmaceutical products measured on 500‐ and 600‐MHz NMR spectrometers. The influence of the most important acquisition (spectral width, transmitter [frequency] offset, number of scans, and time domain) and processing (size of real spectrum, deconvolution, Gaussian window multiplication, and line broadening) parameters for qNMR was automatically investigated. Moderate modification of the majority of the investigated parameters from default instrument settings within evaluated ranges does not significantly affect the trueness and precision of the qNMR. Lite Gaussian window multiplication resulted in accuracy improvement of the qNMR output and is recommended for routine measurements. In general, given that the acquisition and processing parameters were selected based on the presented guidelines, automated qNMR analysis can be employed for reproducible high‐precision concentration measurements in practice.