一种提高SIFT特征匹配正确率的方法

针对尺度不变特征变换(SIFT)特征匹配算法存在计算量大、实时性差、误匹率高的问题,提出一种基于距离比率准则的方法来去除SIFT特征匹配中的错误匹配。传统的方法是采用随机选取一致性(RANSAC)方法选取出正确的匹配对,但是需要通过反复迭代,复杂、耗时并且仍存有部分误匹配的现象。改进后的方法直接通过两条匹配直线斜率的一致性判断,剔除不在斜率范围内的匹配,此方法算法简单,省时高效,从而较大提高了特征匹配的正确率。实验结果表明,通过采用距离比率准则方法具有较高的匹配精度,同时减少了匹配的时间,使实时性得到提高。     

提高五轴数控机床联动精度补偿方法研究

提高五轴数控机床精度的常用方法有几何精度调整、位置误差补偿、RTCP参数补偿。实践证明,这些方法已很难实现联动精度的进一步提高。为有效解决这一问题,突破现有精度调试方法,提出一种新的综合误差补偿方法,分析其补偿原理,建立了数控机床通用误差补偿表达式,通过在CA摆五轴数控机床的应用,使得联动精度提高了0.02mm,验证了该补偿方法的实用性。联动精度的提高同时也使得试切"S"件外形轮廓尺寸公差从±0.05[2]mm提高到±0.03mm。     

A procedure for evaluating the positioning accuracy of reciprocating friction drive systems

The dynamic characteristics of stick-slip motion in reciprocating friction drive systems are investigated under dry contact using a 0.45% carbon steel pair. Based on this dynamic analysis, the stick-slip motion can be eliminated under certain experimental conditions depending upon driver speed, normal load and spring constant. The effects of normal load, driver speed, and spring constant on the positioning accuracy of the reciprocating friction drive system are examined under harmonic oscillation without stick-slip motion. Results show that at very low spring constants, the slip increases with increasing driver speed such that higher normal load has better positioning accuracy or smaller slip than does lower normal load. However, at high spring constants and high normal loads, there exhibits severe wear or peak at two limit positions on the sliding region. As a result, smaller normal loads have a better positioning accuracy than larger normal loads due to severe wear at high normal loads.     

A new method for detection of internal flaws in ferrodielectrics and ferroelectrics

A fundamentally new method for detection of internal flaws in ferrodielectrics and ferroelectrics is substantiated theoretically and experimentally. Traditional acoustic methods are based on the flaw-surface reflection of acoustic waves that occurs owing to abrupt changes in elastic properties at an interface. In contrast, the generation of acoustic waves that occurs upon incidence of EM waves on the flaw surface and that results from abrupt changes in the magnetoelastic and piezoelectric properties forms the basis of the new method.     

Estimating dimensions of flaws with narrow openings along the beam direction

The problem of flaw dimension measurement along the beam direction is discussed from the viewpoint of the measurement accuracy. Practical recommendation are put forth for estimating dimensions of flaws with small openings, such as incomplete fusions in welded joints, along the beam direction using X-ray photographs.     

一种提高螺栓连接强度的有效方法

介绍提高螺栓连接强度的一般方法，提出一种通过增加钢套的简单方法以增大螺栓的柔度，从而提高螺栓连接强度；并从降低螺栓应力幅方面说明该方法能提高螺栓连接强度，运用能量的方法分析其提高螺栓连接抗冲击的原理。     

一种提高模糊控制器稳态控制精度方法的研究

针对模糊控制器本身在消除系统稳态误差时性能比较差,通过对隶属度函数和控制精度之间关系的阐述,提出了一种简单的方法,通过对隶属度函数的调整使控制精度得到提高.仿真结果表明,该方法有较理想的性能、较高的稳态控制精度和广泛的实用性.     

微转换器ADμC812数据采集精度的软件提高方法

使用通常的单片机、A/D转换器等芯片构造一个数据采集系统 ,往往设计周期长 ,成本较高。ADμC81 2微转换器采用高性能的闪速 /电擦除存储器技术和模拟测量技术 ,能灵活的对芯片进行编程 ,大大降低数据采集系统的开发时间和成本。ADμC81 2装有工厂编程的校准系数 ,它在上电时自动下载到ADC ,可以获得很高的精度 ,并且可以采用软件进一步校准 ,从而获得更高的精度 ,确保最佳的ADC性能     

A study of improving overlay accuracy for a stepper in IC manufacture

In integrated circuit manufacture, owing to the limited optical system capability and the possibility of inaccurate mask installation in a stepper, image distortion of the mask pattern projected onto the wafer surface may occur during exposure. In this paper, the overlay distribution geometry was analysed to explore the physical properties of parameters in an overlay accuracy model for a stepper. The multiple linear regression method was also used to analyse overlay models to compare the error parameters obtained by different overlay models and the extent of possible improvements in overlay accuracy. If an appropriate overlay accuracy model can be established to reduce the systematic errors, overlay accuracy can be effectively improved. A multiple linear regression analysis of a simulation study was carried out. The elimination of the systematic error parameters was achieved by finding out the causes of these systematic errors in the overlay and eliminating them. This can reduce the standard deviation of the residual overlay to 19% of the original overlay and thus significantly improve the overlay.     

Computer analysis of the configuration of the magnetic fields of surface discontinuity flaws of finite dimensions in a ferromagnetic plate of finite dimensions by the spatial integral equation method

The spatial integral equation method is used to perform a numerical analysis of the magnetic-field configuration for ferromagnetic objects of finite dimensions having surface discontinuity flaws of arbitrary shape and finite dimensions. The range of application of the software developed for the design of technical devices for inspecting objects of irregular geometry was determined.     

基于提高静压导轨运行精度的方法应用

静压导轨是滑动导轨,通常广泛应用于高精密的加工设备,具有精度高、摩擦系数小、机械效率高等优点。在实际生产过程中,开发、制造等因素在一定程度上影响了静压导轨的精度。针对出现的问题,通过分析改进,提高了静压导轨的运行精度,确保了设备的加工稳定性。     

基于MATLAB的提高测量仪器精度方法研究

针对学生使用的形位公差测量仪器原始误差大、测量精度不高等现状,提出一种基于MATLAB曲面拟合提高测量精度的方法,并给出了具体的数学模型。实验中采用前后对照的实验方法,对同一工件同一位置直线度误差测量,直线度误差值由原来的66.75μm变成34.33μm,精度提高48.6%。实践证明:此方法可以有效减小测量仪器原始误差,并且可以为其他测量仪器消除原始误差提供借鉴。     

A method for evaluating anti-wear lubricant effects

A method is described for estimating the protective (anti-wear) capability of liquid surface-active and inactive media used as lubricating-cooling technological fluids for ‘sparing’ treatment of single crystal surfaces and also for precision mechanical equipment with friction bearings containing single crystals. The method permits the quantitative evaluation of the efficiency of the protective (anti-wear) action of different media based on statistical characteristics of the elasticity-plasticity transition in the subsurface layer of tribopair materials in moving contact.     

A calculation method for evaluating defect parameters in steel

Russian Journal of Nondestructive Testing - This work aims at taking another step forward towards the solution of the inverse problem. A calculation method has been proposed for evaluating the size...     

A decrease in the error of the statistical method for estimating the parameters of flaws in magnetic flaw detection

A method for increasing the accuracy of evaluating the dimensions of flaws during magnetic flaw detection in steel sheets and pipes, which is based on the division of a set of flaws into subclasses using a certain generalized parameter in accordance with the criterion of minimization of an appropriate estimate, is described. The results of applicability of this method to the evaluation of the depths of flaws of the metal loss type are presented.     

Noniterative algorithm for improving the accuracy of a multicolor-light-emitting-diode-based colorimeter

We present a noniterative algorithm to reliably reconstruct the spectral reflectance from discrete reflectance values measured by using multicolor light emitting diodes (LEDs) as probing light sources. The proposed algorithm estimates the spectral reflectance by a linear combination of product functions of the detector's responsivity function and the LEDs' line-shape functions. After introducing suitable correction, the resulting spectral reflectance was found to be free from the spectral-broadening effect due to the finite bandwidth of LED. We analyzed the data for a real sample and found that spectral reflectance with enhanced resolution gives a more accurate prediction in the color measurement.     

A method for estimating the accuracy of quadrature output sensors

The errors of quadrature signal interpolation are analyzed. Quantitative assessment of the noise effect and the quadrature signal distortion is given. A method for estimating the accuracy of quadrature output sensors is proposed.     

Collapse simulation of tubular structures using a finite element limit analysis approach and shell elements

This paper describes the collapse simulation of thin-walled tubular structures using a finite element limit analysis approach and degenerated four-node shell elements. The simulation traces the path of sequential deformation of the structure modelled by considering the strain-hardening effect, which is important for the analysis of collapse behaviour and energy absorption efficiency. The collapse analysis of some square tubes was used to verify the simulation method proposed. Numerical results are compared with experimental observations for sequential collapse loads and deformation modes, showing fairly good coincidence. The collapse analysis of an S-rail was then carried out for sequential collapse loads as well as deformation modes and its results are compared with elasto-plastic analysis results obtained from the explicit dynamic code PAM-CRASH. The energy absorption capacity was studied for a variety of rectangular cross-section aspect ratios. The results show that the energy absorption capacity increases as the height-to-width aspect ratio becomes larger. Results also demonstrate that the finite element limit analysis can predict the plastic collapse load and collapse mode of thin-walled structures efficiently and systematically. The present algorithm with a simple formulation has the advantage of stable convergence, computational efficiency and easy access to strain-hardening materials compared to the incremental rigid–plastic finite element analysis.     

一种提高LVDS可靠性通信方法的研究

为保证远程测试系统中基于低振幅差分信号技术(low voltage differential signaling,LVDS)串行数据传输的可靠性与准确性,提出了一种(7,4)线性分组码.该方法以牺牲一定有效带宽为代价,通过对线上传输的有效数据增加监督码元极大地降低了数据传输的误码率,提高了数据传输的可靠性.采用驱动器CLC001和自适应均衡器CLC012补偿信号在长线传输中的衰减,提高了信号的完整性,保证了数据传输的可靠性.经仿真验证,在总长为100 m的屏蔽双绞线上,该系统以240 Mbit/s的码率实现了串行数据的零误码率传输.