一种新的GVF Snake模型自动初始化轮廓方法

Snake模型已被广泛地应用于数字图像分析和计算机视觉领域.在应用该模型时,初始轮廓不易实现自动化.因此,提出一种改进分水岭算法确定GVF Snake模型初始轮廓的方法.该方法先采用改进的分水岭算法对图像进行预分割,再将得到的边缘作为GVF Snake模型的初始边缘轮廓.实验结果证明:该方法极大地减少了GVF迭代次数,提高了轮廓收敛的速度,并保留了GVF Snake模型的所有优点,为Snake模型自动初始化轮廓提供了一种有效的方法.     

基于短时傅里叶变换的光时域反射计(OTDR)事件分析

论文研究了基于单边指数窗的短时傅里叶变换(又称Gabor变换)的光时域反射计(OTDR)事件分析算法。在分析信号模型中噪声的产生原理的基础上,简化了事件检测算法。提出应用Gabor变换和二元信号检测理论对OTDR数据进行事件检测。该算法在保证准确检测事件信息的基础上,提高了OTDR远端事件检测的能力,提高了算法的速度和效率。对实测数据的处理表明,该算法可以有效定位光纤中的事件信息,具有较强的应用价值。     

A study on CFD data compression using hybrid supercompact wavelets

A hybrid method with supercompact multiwavelets is suggested as an efficient and practical method to compress CFD dataset. Supercompact multiwavelets provide various advantages such as compact support and orthogonality in CFD data compression. The compactness is a crucial condition for approximated representation of CFD data to avoid unnecessary interaction between remotely spaced data across various singularities such as shock and vortices. But the supercompact multiwavelet method has to fit the CFD grid size to a product of integer and power of two,m _X ². To resolve this problem, the hybrid method with combination of 3, 2 and 1 dimensional version of wavelets is studied. With the hybrid method, any arbitrary size can be handled without any shrinkage or expansion of the original problem. The presented method allows high data compression ratio for fluid simulation data. Several numerical tests substantiate large data compression ratios for flow field simulation successfully.     

基于图像数据的平面轮廓加工CAM技术

提出了简化的霍夫变换HT(HoughTransform)方法和适应性包容盒AIB(AdaptiveIncludingBox)原理.适应性包容盒AIB的大小可以通过加工精度控制,零件轮廓的图像数据能够转化为分段的直线,进而生成零件轮廓加工的数控代码.     

A new method to determine the number of experimental data using statistical modeling methods

For analyzing the statistical performance of physical systems, statistical characteristics of physical parameters such as material properties need to be estimated by collecting experimental data. For accurate statistical modeling, many such experiments may be required, but data are usually quite limited owing to the cost and time constraints of experiments. In this study, a new method for determining a reasonable number of experimental data is proposed using an area metric, after obtaining statistical models using the information on the underlying distribution, the Sequential statistical modeling (SSM) approach, and the Kernel density estimation (KDE) approach. The area metric is used as a convergence criterion to determine the necessary and sufficient number of experimental data to be acquired. The proposed method is validated in simulations, using different statistical modeling methods, different true models, and different convergence criteria. An example data set with 29 data describing the fatigue strength coefficient of SAE 950X is used for demonstrating the performance of the obtained statistical models that use a pre-determined number of experimental data in predicting the probability of failure for a target fatigue life.

     

一种基于傅里叶变换的双级匹配掌纹识别系统

提出了一种基于傅里叶变换的双级匹配掌纹识别系统。首先提取掌面的最大内圆作为识别的有效区域,利用该内圆的半径进行第一级匹配。然后对提取出的区域进行傅里叶变换,统计出频域图像的能量分布模式进行第二级匹配。实验表明这种双级匹配的识别方法实现简单,操作方便,同时具有较快的反应速度和较低的误识率。     

Wave field restoration using three-dimensional Fourier filtering method.

A wave field restoration method in transmission electron microscopy (TEM) was mathematically derived based on a three-dimensional (3D) image formation theory. Wave field restoration using this method together with spherical aberration correction was experimentally confirmed in through-focus images of amorphous tungsten thin film, and the resolution of the reconstructed phase image was successfully improved from the Scherzer resolution limit to the information limit. In an application of this method to a crystalline sample, the surface structure of Au(110) was observed in a profile-imaging mode. The processed phase image showed quantitatively the atomic relaxation of the topmost layer.     

傅里叶变换红外光谱学方法用于气体定量分析

本文将傅里叶变换红外光谱学方法运用到CO和CO2气体的定量分析中，通过红外光谱仪系统测量得到待测气体的红外透过率光谱，采用非线性最小二乘拟合算法基于红外标准数据库HITRAN中的CO和CO2气体光谱数据分别对测量得到的光谱进行拟合，得出待测气体的浓度。对于密闭样品池CO透过率光谱测量分析结果的相对误差小于5％；对于近距离大气开放光路CO2透过率光谱测量分析的相对误差小于1％，实验结果表明基于FTIR方法进行气体浓度信息的定量分析是可行的。     

基于SVDD的分类新方法

为了实现对大规模数据集的分类,提出了一种基于支持向量数据描述（SVDD）的新算法。把两类的支持向量机（SVM）问题转化为软间隔的SVDD问题,然后用核集迭代的方法得到了近似最优解。实验结果表明,提出的算法与传统的SVM方法预测精度相当,但是训练时间和测试时间更短,产生的支持向量数也更少,适用于大规模数据的分类。     

一种稳健的高光谱图像压缩方法

提出了一种稳健的高光谱图像压缩技术。估计原始数据的虚拟维数,继而采用一种非监督的正交子空间投影技术提取端元矢量,构成矩阵作为快速独立分量分析的初始化混合矩阵,进行ICA变换。对变换后的独立分量图,采用二维分层树集合分裂进一步压缩编码。计算机仿真结果表明,该压缩算法在取得40∶1压缩比的同时,能有效地保持数据,尤其是小目标和异常的谱向特征,是一种高效的三维数据压缩方法,可有效地应用在小目标检测及异常检测中。     

一种改进的轮廓跟踪算法

针对轮廓数据在基于工业CT图像快速反求工程中的重要作用,详细分析了Freeman链码跟踪法,根据反求工程的具体要求,提出了一种改进的轮廓跟踪法和基于轮廓跟踪数据的优化方法.对具体的ICT图像进行了轮廓跟踪,抗干扰和数据优化等分析处理,其结果不但能正确跟踪到所需轮廓数据,而且明显精简了跟踪结果的数据量,对快速原型生产具有重要意义.     

用改进的快速傅里叶变换法评估一维标准样板线节距数据

由于利用快速傅里叶变换(FFT)法定量测量标准样板的平均节距时会因为频谱分辨率Δf有限而在数据分析过程中丢失信息,故本文对常用的FFT法进行了改进。首先在FFT的频谱中得到最大振幅值对应的频率分量fmax,使用二分法缩小fmax附近的频率范围;然后利用连续傅里叶变换法寻找更大的振幅值与更精确的f′max以求得更接近真实值的线节距。为了验证该方法的可行性,利用Matlab软件仿真被测标准样板(TGD1)的轮廓图,比较了FFT与改进后的FFT对不同扫描长度下数据的评估结果,并展示了改进FFT法评估时部分数据的运算过程。同时,通过计量型原子力显微镜实测20μm×2μm内的TGD1形貌。两种评估法下的数据比对结果显示:改进FFT评估后的线节距为(277.84±0.39)nm,符合标称值(278±1)nm,验证了改进后FFT方法对线节距求解的准确性与合理性。     

一种新型压缩机构及其产品开发

介绍了一种新型容积式压缩机构及利用这种新机构在气体压缩机和油泵等流体机械产品上的应用及开发情况，并与其它机构的产品进行了比较、研究。     

A new velocity estimation method using spectral identification of noise

In all measurement techniques one seeks accuracy and precision. In ultrasonic Doppler velocimetry, those qualities strongly depend on signal to noise ratio of the Doppler signal and on the performance of the velocity estimator. The most widely used estimation method in ultrasonic coherent Doppler velocimetry is the pulse pair method. Its success is due to the computation efficiency of the algorithm combined to an unbiased estimator. Unfortunately, for a wide range of experimental fluid flows, the pulse pair estimation is less efficient, especially for clear water or concentrated mud where the signal to noise ratio can be very low, or for highly turbulent flows where the Doppler signal has a broad spectrum. Our approach is based on the treatment of the Doppler spectral information. It uses a simple parametric identification inspired by theoretical models and experimental observations. It acts through noise subtraction and subsequent cutting. Thus, we have developed a fast velocity estimation algorithm superior to the pulse pair one in terms of accuracy. The robustness of the method was evaluated by adding different levels of white Gaussian noise to an experimental Doppler signal. The results demonstrate an increase of noise immunity up to one decade compared to the pulse pair method.     

A new method of controlling cavity-induced pressure oscillations using sub-cavity

A new passive control technique of cavity-induced pressure oscillations has been investigated numerically for a supersonic two-dimensional flow over open rectangular cavities at Mac h number 1.83 at the cavity entrance. A subcavity on the front wall of the cavity covered by a flat plate was evaluated for the effectiveness of controlling cavity-induced acoustic oscillations. The results showed that sub-cavity is very effective in reducing cavity-induced pressure oscillations. The results also showed that the resultant amount of attenuation of cavity-induced pressure oscillations was dependent on the length and thickness of the flat plate, and also on the depth of the sub-cavity used as an oscillation suppressor.     

A new data analysis method to determine pesticide concentrations by cyclic voltammetry

Determining pesticide concentration is very important for safe agricultural products. In this study, a new data analysis method was developed to determine the concentration of Carbofuran pesticides by cyclic voltammetry. The peak current and two specific area indexes calculated from the cyclic voltammetry curves served as the dependent variables. The concentrations of Carbofuran pesticides were used as the independent variables. The inverse calibration equation was used to evaluate the predictive performance. The results indicated that the numeric values of the cyclic voltammogram area ranging from 0.5 to 0.9 V were the best index value of pesticide concentration. The precision of this index was better than that of others two indexes. The accuracy of the determination technique with the calibration equation was within 5 ppm by the adequate calibration equation.     

High order level contour reconstruction method

Complex interfacial physics arising from geometric curvature associated with surface tension as well as phase transformation make it a formidable task to design an accurate, reliable, and yet simple method for direct computation of multiphase flows. Hybrid methods mixing conventional, Volume-of-Fluid, Level Set, Phase Field, and Front Tracking methods have recently become popular in an attempt to overcome the shortcomings of each method alone. We developed the Level Contour Reconstruction Method (LCRM) as part of a hybrid method for treating the complex interface geometry associated with general three-dimensional multiphase flows. The main idea in that work focused on a simple and robust algorithm especially suited for dynamic interfaces in the three-dimensional case by combining characteristics of both Front Tracking and Level Set methods. In this article we describe a modification to the LCRM which introduces a high order interpolation kernel during the course of the interface reconstruction along with a new hybrid surface tension formulation. With this we can essentially eliminate any mass redistribution between regions of differing curvature and reconstruct the interface accurately and smoothly. The improvement with high order reconstruction is also noticeable vis a vis spurious currents which are further decreased by two orders of magnitude over the previous linear reconstruction method. Moreover, there is no disturbance concurrent with reconstruction and the solution fidelity is not influenced by the reconstruction time step. This High Order Level Contour Reconstruction Method retains the simplicity of the original LCRM and avoids complicated interface smoothing procedures.     

基于重复子向量的测试数据压缩算法

随着微电子技术的快速发展,系统芯片SoC的集成度越来越高,所需的测试数据量成几何级数增长.针对这一问题,本文提出了一种有效的测试数据压缩算法--基于重复子向量的测试数据压缩算法.该算法适用于多扫描链设计的IP芯核,应用过程中不需要芯核内部结构信息.该方法针对测试集中大量存在的重复子向量进行压缩,能够有效提高压缩效率.理论分析和实验数据表明,基于重复子向量的测试数据压缩算法相对于同类压缩方法能够大幅度提高压缩效率、降低测试成本.     

基于改进的最小描述长度小波变换数据压缩

针对电厂数据种类多、数据量大,而实际存储空间有限的情况,提出了一种基于改进最小描述长度（MDL）的小波变换数据压缩方法.该方法根据用户关注的信息特征,对各频段的信息加上不同的权值,以改进最小描述长度函数为准则.通过对电厂负荷数据进行压缩和解压实验,不但选择了合适的小波基函数,而且结果显示出改进MDL方法能够获得优于MDL方法的数据压缩比和重构精度.     

A real-time data compression algorithm for gear fault signals

A flexible, stable and controllable real-time algorithm of Auto-Regressive and Moving Average based on Swing Door Trending (ARMA-SDT) is proposed for the compression of impact-type signals in gear fault detection systems. The Auto-Regressive and Moving Average (ARMA) model is used to predict the variation trend of signal features. To guarantee the adaptability, an empirical equation is proposed to calculate the compression threshold of the Swing Door Trending (SDT). Based on the empirical equation and prediction results, dynamic self-regulation of compression threshold is realized, and the compression error always stays around a preconfigured value. Moreover, an experimental setup and an engineering solution are proposed to verify the usefulness, reliability, and stability of the proposed ARMA-SDT algorithm in data compression.