基于高斯混合模型的阴影消除算法

关于阴影消除是智能监控系统的重要组成部分,且对后续的识别和跟踪有很大的影响。根据改进的高斯混合模型检测出目标,用高斯混合阴影模型来检测阴影,并用高斯平滑滤波处理检测的结果。然后动态选择合适的阈值对图像做二值化处理,进一步对结果进行形态学滤波。经过整个过程的处理,得到了较为完整的轮廓,很好的消除了阴影,提高后续的识别和跟踪的准确性和可靠性。实验表明,算法能较好地消除阴影,且具有较好的实时性和鲁棒性。     

Type-2 fuzzy Gaussian mixture models

This paper presents a new extension of Gaussian mixture models (GMMs) based on type-2 fuzzy sets (T2 FSs) referred to as T2 FGMMs. The estimated parameters of the GMM may not accurately reflect the underlying distributions of the observations because of insufficient and noisy data in real-world problems. By three-dimensional membership functions of T2 FSs, T2 FGMMs use footprint of uncertainty (FOU) as well as interval secondary membership functions to handle GMMs uncertain mean vector or uncertain covariance matrix, and thus GMMs parameters vary anywhere in an interval with uniform possibilities. As a result, the likelihood of the T2 FGMM becomes an interval rather than a precise real number to account for GMMs uncertainty. These interval likelihoods are then processed by the generalized linear model (GLM) for classification decision-making. In this paper we focus on the role of the FOU in pattern classification. Multi-category classification on different data sets from UCI repository shows that T2 FGMMs are consistently as good as or better than GMMs in case of insufficient training data, and are also insensitive to different areas of the FOU. Based on T2 FGMMs, we extend hidden Markov models (HMMs) to type-2 fuzzy HMMs (T2 FHMMs). Phoneme classification in the babble noise shows that T2 FHMMs outperform classical HMMs in terms of the robustness and classification rate. We also find that the larger area of the FOU in T2 FHMMs with uncertain mean vectors performs better in classification when the signal-to-noise ratio is lower.     

自适应背景提取算法的比较

介绍几种背景自适应提取算法的原理,分析算法在交通视频检测中的仿真效果.依据交通场景提取中实时性、准确性等关键技术指标,采用基于"真理图"的量化评估方法,在多种复杂交通场景中利用recall,precision评价参数对背景提取算法的效果进行比较分析,明确被评价算法的性能特点与适用的交通场景.     

组合ARMA与SVR模型的时间序列预测

经典的ARMA模型常用于平稳时间序列的预测,而对于自然界绝大部分的非平稳序列一般采用确定性时序分析和随机时序分析.确定性时序分析对随机性信息浪费严重,而随机时序分析经过差分平稳序列后又回归到ARMA模型.本文利用在充分ARMA模型拟合后的残差序列进行支持向量回归(SVR)拟合,进而对原序列进行组合预测,比起单一模型的拟合及预测,该组合有效地提高了预测精度.     

氨水溶液汽液界面微观结构的分子模拟

为研究吸收式制冷机常用工质对氨水溶液的微观形态,采用分子动力学方法:蛙跳法积分运动方程,Ewald加和方法计算库仑相互作用,分别对不同摩尔浓度的氨水溶液的汽液界面的微观结构、密度分布、界面厚度、界面张力进行r分析与计算,并拟合出界面厚度随氨水溶液摩尔浓度变化的线性关系式.结果表明:氨分子比水分子更容易吸附于汽液界面处,随着氨水溶液摩尔浓度的增加,汽液界面厚度增加,汽液界面张力减小,变化趋势与实验一致.界面张力的计算值从计算结果的量级和数量上看,与实验值接近.     

烃类混合物液相粘度的预测

介绍了计算粘度的Przezdziecki-Sridhar模型,Sastri-Rao模型,PRη模型和Dymond-Assael硬球模型,并分别用它们计算庚烷-辛烷-壬烷,庚烷-辛烷-己基苯,庚烷-壬烷-己基苯和辛烷-壬烷-己基苯4个三元体系的从293.15 T到313.15 T温度的粘度.将计算结果与实验值比较,发现用Sastri-Rao模型,Dymond-Assael硬球模型和PRμ模型误差较小,而用Przezdziecki-Sridhar模型误差较大.     

MBS对PVC的增韧作用及其机理的探讨

本文通过扫描电镜观察了ＰＶＣ／ＭＢＳ共混物银纹的发展和冲击断面形貌,并测定了银纹体的密度变化。采用光散射法分析银纹和剪切带比例。合成不同玻璃化温度的ＭＢＳ考察它对ＰＶＣ增韧的影响,进行有关ＭＢＳ对ＰＶＣ增韧机理的探讨。     

Correlation and extrapolation scheme for the composition and temperature dependence of viscosity of binary gaseous mixtures: Carbon dioxide + ethane

Experimental viscosity data of ethane, carbon dioxide, and three mole fractions of the binary system carbon dioxide + ethane in the temperature range 293.15<T633.15 K and in the density range 0.010.05 mol·L^–1 reported earlier were evaluated simultaneously to find out a useful correlation and extrapolation scheme for the viscosity of binary systems in the range of moderate densities. A procedure based on the ideas of the modified Enskog theory has been found to give the best results. Dependent on temperature, the collision diameters related to the equilibrium radial distribution function at contact are fitted to viscosity values of the pure substances and of at least one mixture. The results are compared with experimental data from the literature. A recommendation is given concerning the density range in which the first density contribution to the viscosity coefficient of the system carbon dioxide + ethane is sufficient to be included.Paper presented at the Twelfth Symposium on Thermophysical Properties, June 19–24, 1994, Boulder, Colorado, U.S.A.     

基于机器学习的配电网异常缺失数据动态清洗方法

针对传统配电网数据清洗过程中异常数据判断阈值需要人为设定、缺失数据填补效率不佳的局限性,提出基于机器学习的配电网异常缺失数据一体化动态清洗方法。首先,基于局部异常因子检测算法和高斯混合模型,提出一种异常数据动态检测改进算法,实现配电网异常数据阈值的准确自动选择。其次,基于随机森林算法与最小二乘回归法,提出一种配电网缺失数据动态填补算法。根据缺失数据时间长度自适应优化填补算法,在保证数据填补精度的同时降低计算时间。在此基础上,通过异常数据检测和缺失数据填补共同构建一体化动态清洗架构。采用湖南某地区配电网数据进行实例验证,结果表明所提方法可实现异常辨识阈值准确自动选择,有效检测配电网异常数据,并且实现缺失数据填补精度与速度的平衡,具有较好的工程应用价值。     

多源信息融合的航空发动机异常检测方法研究

针对航空发动机中单参数表征性能不全面,且易受外部环境和飞行工况等因素影响的问题,提出一种基于谱回归和高斯混合模型的多源信息融合的健康因子构建方法。选取记录完整并与发动机健康性能关系密切的参数,以同架飞机左右发动机的性能参数差异值作为数据源,通过谱回归进行特征降维,采用高斯混合模型构建正常状态模型,再以基于贝叶斯推断的距离表征测试数据与高斯混合模型的全局距离判别发动机的异常状态。通过两个航空发动机异常事件案例中真实QAR数据进行验证,结果表明所提方法相较于航空公司能够更有效评估航空发动机的健康状态,提前识别出发动机的异常点,预留出足够的时间为发动机制定可靠的维修计划,提高飞机的安全性与经济性。