基于像素分组的帧内预测快速算法

为了降低HEVC帧内编码的复杂度,基于HM 12.0提出了一种帧内预测快速算法.首先将35种预测方向分成5组,然后基于像素提取PU块的梯度方向信息并匹配到其中一组,最后在RDO运算时根据统计信息适当降低了候选模式的数量.实验结果表明,与JCTVC最新源码HM 12.0相比,该算法平均能够节省20％的编码时间,峰值信噪比和码率变化不大.     

The search for magnetic monopoles in magnetite from North China

Using the 27 kG magnetic field produced by a superconducting solenoid we have searched for trapped magnetic monopoles in a magnetite sample from North China. We hvve scanned the nuclear emulsion both for low ionizing tracks and heavily ionizing tracks, and polycarbonate track detectors for heavily ionizing tracks of magnetic monopoles. We obtained a 90% confidence level upper limit of 1.8×10⁻²⁶ monopoles per nucleon in the magnetite sample.     

Optimization of the heavy nuclei track etching in olivine crystals from meteorites

The investigation of the track etching velocity dependence for nuclei with Z50 on pH of the etching solution in the olivine crystals from Marjalahti and Eagle Station meteorites, containing galactic cosmic-ray nuclear tracks and also the crystals from Marjalahti meteorite, calibrated at the UNILAC (GSI, Darmstadt) with 14.17 MeV/N ²⁰⁸Pb nuclei, were performed.

The pH of the etching solution were changed from 7.8 to 9.4. It is shown that the optimal track etch condition were realised at pH of etching solution in the interval from 8.4 to 8.8.

The results of the Galactic cosmic-ray nuclei tracks with Z83 revealed in olivine crystals are presented. They were additionally irradiated at 90% to the surface with accelerated 14 MeV/N ¹³²Xe nuclei.     

全三维电离粒子有源像素探测器优化仿真

提出了一种双外延高能离子注入的单片集成有源像素探测器的传感器结构,以提升传感器对电荷的收集性能和辐射加固,并进行了三维工艺模拟和物理级器件仿真计算.研究结果表明,所提出的传感器结构改善了内部电场和电势的分布,且目标电极的电荷收集效率提高70%,电荷收集时间减少64%.此外,当等效中子辐射流通量在1012—1015cm-2范围内时,所提出的传感器结构比标准传感器结构有更高的电荷收集能力.     

Sub-pixel dimensional measurement with Logistic edge model

To separate substandard workpieces from products, sub-pixel edge location with Logistic edge model is proposed for dimensional measurement. In the first step, workpiece rotation is corrected with high precision image rotation to reduce the measurement error and simplify the edge model. Secondly, Logistic edge model is utilized for fitting the discrete edge data to locate the sub-pixel workpiece edge and model parameters are estimated through the objective function in terms of the difference between assumed model and real image data. Finally, an edge pixel compensation method is presented with edge pixel compensation value, which weakens the effect of edge transition zone effectively. Experimental results demonstrate that the proposed method is effective for dimensional measurement with the relative error less than 0.05%.     

Correction for Shift-variant Characteristics of Single Photon Emission Computed Tomography Acquisition System Using Sensitivity Functions

The single photon emission computed tomography (SPECT) imaging system has shift-variant characteristics due to non-uniform attenuation of gamma-ray, collimator design, scattered photons, etc. In order to provide quantitatively accurate SPECT images, these characteristics should be compensated in the reconstruction. This paper presents a new method to correct the shift-variant characteristics, which is based on a continuous-discrete mapping model and filtered backprojection (FBP) method, in which the projection data are assumed to be acquired by narrow ray sum beams in the FBP method and the assumed data set is expressed as a linear combination of the actual projection data. Narrow ray sum beams are approximated by a weighted sum of the original sensitivity functions. Thus, at the reconstruction the projection data are first modified using an approximation and the FBP method is then applied to the corrected projection data and a SPECT image is reconstructed. We further propose a technique that requires the inversion of smaller matrices than the conventional algebraic method; the amount of calculation and memory space become smaller and the stability of the calculation is greatly improved as well. The results of the numerical simulations are also demonstrated.     

电子清管器发射机系统的改进设计

现有的以555芯片产生间断超低频（23Hz左右）电磁脉冲的电子清管器发射机系统会带来频率漂移、各组振荡问的相位关系随机不恒定及静态功耗偏大的问题。为此利用89c2051编程产生间断的超低频脉冲信号,解决了频率漂移及振荡间的相位关系随机的问题。通过采取配置合适的晶振和电源电压措施,并在软件设计部分充分利用单片机休眠模式的低功耗特性,解决了功耗问题。经测试频率稳定在（233-0．01）Hz,振荡间的相位关系恒定。有效工作时间由60h提高至100h。     

遮挡情况下目标跟踪算法综述

目标跟踪是机器视觉的核心问题之一。由于被跟踪目标本身特征的多样性,所处环境的复杂性,使得目标跟踪成为一个极富挑战性的课题。特别是跟踪过程中出现的遮挡问题,一直是目标跟踪中的难点。如何有效处理遮挡,尤其是严重遮挡和全部遮挡,成为限制目标跟踪算法鲁棒性的关键因素。本文对国内外现有的遮挡情况下的目标跟踪算法进行了总结和分类,阐述了各种算法的基本思想,简要介绍了其代表性的算法,指出了算法的适用场合,分析了算法的优缺点。文章最后对遮挡情况下的目标跟踪做了必要的讨论。     

钕铁硼在高温超导磁悬浮技术实验中的应用

将新材料钕铁硼作为磁导轨引入到高温超导磁悬浮技术实验中,介绍了高温超导悬浮技术实验的基本原理、仪器及钕铁硼对实验的改进.     

梯度算子选择对基于梯度的亚像素位移算法的影响

基于梯度的亚像素位移算法中不同的灰度梯度计算方法对计算结果的影响进行了研究,列出了几种常用灰度梯度计算方法,并用对相邻5像素点的离散灰度进行三次样条插值的方法推导出两种灰度梯度算子。然后对计算机模拟生成的散斑图像对用各种梯度算子计算得出的亚像素位移与预先设定的真实值做了比较,给出了相应的均值误差和均值相对误差。最后的真实试件的刚体平移和单向拉伸实验结果表明,Barron算子是所列的几种梯度算法中最为精确、稳定的梯度算子。