基于Single-Pass的网络舆情热点发现算法

考虑网络事件的时间距离,基于半结构化网页中不同位置特征项重要程度的不同,提出改进的single-pass文本聚类算法single-pass*,优势在于对Web文本不同位置特征项的加权处理,仅需计算新文档与同类别种子文档间的相似度。实验结果表明,相比single-pass,改进算法极大减少了漏检率和错检率,降低了由于新文本流内文档进行相似度计算导致系统性能的下降,平均提高Web文本聚类效率40%。将聚类后的Web文本应用于网络舆情分析,进行主题关注度分析和话题热度特性分析。     

仿射变换在压缩感知跟踪中的应用

针对跟踪中目标尺度变化和旋转问题,将仿射变换和应用到压缩感知跟踪中.首先,以上一帧的跟踪结果为均值,以一定的标准差按照高斯分布,随机生成不同尺度和旋转角度的候选框;然后,通过仿射变换将其转换至直角坐标系中,通过多尺度滤波得到目标在不同尺度下的高维特征向量,采用压缩矩阵将高维特征向量降维至低维空间;最后,将低维特征向量通过贝叶斯分类器选取具有最大响应的候选位置作为目标的跟踪位置.在此基础上分别提取正负样本来更新分类器参数,从而实现持续稳定的跟踪.实验结果表明,该算法能够较好地解决压缩感知跟踪中的目标旋转和尺度变化问题.     

Robust adaptive threshold speech endpoint detection method

Due to the fact that traditional Speech Endpoint Detection methods' performance degrads greatly in a low signal-to-noise ratio and nonstationary noise, a novel robust adatpive threshold endpoint detection method is proposed. First of all, the LSFM parameter is employed as the distinctive feature and the Burg spectrum estimation is applied to figure out the power spectrum, which can enhance the discriminative ability in classifying speech signals and noise, compared with the traditional speech features. Furthermore, an adaptive threshold based on the Bayes estimation criterion is involved in the final judgment, which overcomes the defect of the fixed threshold in adaptability and improves the detection performance to a greater degree. Simulation results show that compared with the traditional feature-based Speech Endpoint Detection methods, the accuracy of the proposed method has a high accuracy rate, which proves that the new method has a better robust performance in a low SNR and nonstationary noise.     

Wonderful events automatic detection and location of the HCRF and webcast text

Based on the powerful function of the hidden conditional random fields (HCRF) model in the expression and identification of semantic events and combining the webcast text information, a new framework for wonderful events automatic detection is put forward. Firstly, by analyzing and processing the webcast text, keyword combinations corresponding to each exciting event are obtained. Secondly, by classifying the webcast text to be detected, the happening time labels of each wonderful event are obtained. Thirdly, an HCRF model for semantic shot annotation is built to realize the semantic annotation of multiple types of semantic shots simultaneously, and the semantic shot sequence of the video clip is obtained. Finally, combining the multi-modal semantic clues, an HCRF model for the wonderful events detection and localization is effectively built in the case of small-scale training samples. Based on the mapping relationship among video low-level features, the multi-modal semantic clues and the wonderful semantic events, the inherent patterns of the wonderful events are excavated deeply in the multiple dimensions of the semantic structure, and then the wonderful events automatic detection, localization and segmentation are precisely achieved. Experiments show the effectiveness of this model.     

一种基于SDN的在线流量异常检测方法

基于软件定义网络的集中管控平面,提出了一种在线流量异常检测方法.首先在控制器上在线获取OpenFlow交换机的流表信息,并构造整个网络的流量矩阵与样本熵矩阵进行组合,然后采用主成分分析方法检测异常流量.实验结果表明,相比于传统网络中利用主成分分析方法分别单独处理离线的流量矩阵或样本熵矩阵的方法,在线流量异常检测方法实现和处理方式简单、有效,异常流量能够得到快速隔离,是基于软件定义网络的一种轻量级在线流量异常检测方法.     

Adaptive key SURF feature extraction and application in unmanned vehicle dynamic object recognition

A new method based on adaptive Hessian matrix threshold of finding key SRUF( speeded up robust features) features is proposed and is applied to an unmanned vehicle for its dynamic object recognition and guided navigation. First,the object recognition algorithm based on SURF feature matching for unmanned vehicle guided navigation is introduced. Then,the standard local invariant feature extraction algorithm SRUF is analyzed,the Hessian Metrix is especially discussed,and a method of adaptive Hessian threshold is proposed which is based on correct matching point pairs threshold feedback under a close loop frame. At last,different dynamic object recognition experiments under different weather light conditions are discussed. The experimental result shows that the key SURF feature abstract algorithm and the dynamic object recognition method can be used for unmanned vehicle systems.     

基于BP神经网络的高精度基波频率检测方法研究

电力系统谐波检测过程对基波频率的检测精度要求越来越高。针对这一需求,在对比了各种常用基波检测方法后,提出一种基于BP(Back Propagation)神经网络的高精度基波频率检测方法。既通过充分有效地学习和训练,构建出了合理的BP神经网络模型,并利用Matlab软件对其进行仿真验证。结果表明,算法对频率测量精度达10-3~10-5,并且任一频率(50 Hz除外)BP神经网络算法的频率测量都比加汉宁窗的FFT算法高2~3个数量级。     

基于FDS地铁直线盾构隧道火灾温度探测模拟研究

以北京地铁运行的盾构型隧道为研究对象,利用FDS(Fire Dynamics Simulator)火灾动力学模拟软件进行计算.在不同工况下,对不同热释放速率的火源分别进行探测模拟.通过计算结果,总结在不同位置温度随时间的上升规律.分析不同位置在不同时间段的危害性,为安全逃生提供参考.     

基于分布估计算法的人脸Haar特征选择

通过研究人脸检测算法中Ada Boost算法,针对算法中的haar特征维数过高、训练耗时过长,检测效率过低等问题.提出基于分布估计算法(Estimation of Distribution Algorithm,EDA)的人脸haar特征选择人脸检测.EDA采用类内类间比作为适应度函数,通过统计学习的手段建立解空间内个体分布的概率模型,然后对概率模型随机采样产生新的群体,进行反复计算,实现群体的进化,最终得到全局最优解,以此来实现haar特征选择.实验结果表明:检测率(DR)与误检率(FDR)优于传统算法,而且检测速度得到了提升.     

燃气轮机排气温度异常检测及诊断

针对燃气轮机运行过程中的健康维护问题,提出了一种基于模糊C均值聚类的燃气轮机排气温度异常检测方法。以某型工业燃气轮机为例,采用模糊C均值聚类算法对排气温度的监测数据进行了聚类分析,得到了燃气轮机不同运行状态下的排气温度特征模式,并在此基础上实现了燃气轮机排气温度异常状态下的故障诊断分析。研究结果表明,燃气轮机甩负荷及其热通道部件的损坏失效均能够对排气温度产生不同程度的影响,模糊C均值聚类算法可以有效实现燃气轮机排气温度的异常检测,为燃气轮机性能退化预测及故障诊断提供决策参考。