视频监控系统中的概率模型单目标跟踪框架

针对视频监控的特点与跟踪目标的强机动性,提出了一种新的基于概率模型的目标跟踪框架,从目标表观模型、系统动态模型以及系统观测模型3个方面对当前标准的粒子滤波目标跟踪方法进行了改进。首先,考虑人眼细胞的分布特点,基于人眼分布结构建立目标表观模型来提高跟踪系统抵抗局部遮挡的能力;然后,建立基于自适应目标运动的系统动态模型,提高跟踪算法对快速机动目标的鲁棒性;最后,采用实时更新的系统观测模型,有效避免目标在遇到遮挡、光照变化、剧烈变形等情况下发生的跟踪漂移现象。实验结果表明,本文算法的正确跟踪率可达98%;平均跟踪误差小于6个像元。实验证明本文算法在保证系统跟踪精度要求的同时,具有计算量小、抗干扰能力强等特点。     

非连续路段下移动机器人的轨迹跟踪

针对轮式移动机器人在非连续路段下的轨迹跟踪问题,对移动机器人数学模型和道路模型的建立、轨迹信息的获取以及控制器的设计进行了研究。首先分析了移动机器人的运动学方程及对摄像机的标定,通过建立移动机器人的直线轨迹跟踪模型,并用二值化方法对非连续路段进行了图像处理,提取了图像中非连续路段信息,实现了移动机器人在轨迹跟踪过程中运动的可靠性;然后基于Lyapunov函数设计了渐进稳定的轨迹跟踪控制器,使移动机器人能够在连续路段和非连续路段跟踪确定路径;最后通过Matlab进行了仿真。仿真结果表明,所设计的控制器和算法能使移动机器人的跟踪误差快速收敛于零,轨迹跟踪效果良好,适用于移动机器人对非连续路段的轨迹跟踪控制。     

一种改进的Canny自适应边缘检测算法

针对原始的Canny边缘检测算法高低阈值不利于选择的问题,提出一种改进的Canny自适应边缘检测算法。该方法采用3×3的Sobel边缘算子模板代替2×2邻域内求有限差分来计算梯度幅值;利用梯度直方图来选取高低阈值,提高了算法的自适应性;在边缘的连接上采用边界追踪的方法,来保证边缘的连续性和唯一性。以20m m量块图像边缘检测为例进行了实验。结果表明,该算法在保证实时性的同时,具有更好的检测精度和准确度,细化图像边缘和抑制伪边缘噪声方面的性能也得到了提高,是一种有益的边缘检测改善方法。     

软煤层工作面煤尘分布规律及防治技术

粉尘是影响工作面安全、高效开采的重要因素,尤其对于软煤层工作面而言,粉尘治理是国内外煤炭企业遇到的普遍难题。以21031软煤层工作面为研究对象,分析了该工作面产尘机理,并通过现场实测掌握了该工作面粉尘分布特征和规律。提出采用KHCG500型尘源跟踪降尘系统来治理该工作面粉尘,结果表明,该降尘系统可有效降低21031工作面粉尘,指出应辅以传统的煤层注水、优化通风等粉尘治理措施来综合治理工作面粉尘。     

差分GPS定位技术在列车实时动态追踪系统中的应用

本文介绍了差分GPS定位技术原理,针对矿区列车实时动态追踪系统中使用的位置差分技术和具有多个基准站的局域差分技术进行了分析和探讨。     

Global maximum power point tracking based on new extremum seeking control scheme

A new perturbed‐based extremum seeking control (PESC) scheme is proposed in this paper to track the global maximum power point (GMPP). The PESC scheme has two control loops based on power of the photovoltaic (PV) array: the first loop operates as usually to track the maximum power point and the second sweeps all local MPPs to locate the GMPP. Once the GMPP is located based on its uniqueness (after the PV pattern is quickly scanned many times, depending on the PV pattern's profile), the GMPP is accurately tracked based on first control loop. The used PV patterns have the profile of the PV power characteristics obtained for PV array under partially shaded conditions (PSCs). This PESC scheme is proposed to track the GMPP in the PV applications, but also in other multimodal problems from industry, being a good motif to revive the specialists' interest for the extremum seeking control field. The results obtained here are very promising for both search speed and tracking accuracy performances of the GMPP under different PSCs simulated on the PV array. Thus, the energy efficiency of PV array controlled with the proposed PESC scheme will increase with more than 1.2% in comparison with that obtained with the other MPP algorithms because of better performance shown by this PESC scheme. A 99.6% tracking accuracy is obtained here in comparison with a maximum 98.4% tracking accuracy reported in the literature. Furthermore, 100% hit and high search speed are obtained here for the GMPP localization. Copyright © 2015 John Wiley & Sons, Ltd.     

强跟踪自适应CKF及其在动力定位中应用

针对容积卡尔曼滤波(cubature Kalman filter,CKF)在动力定位系统存在不良测量及船舶运动模型不确定时滤波精确度下降甚至发散的问题,提出一种强跟踪自适应CKF算法。利用新息协方差匹配原理,建立对系统的不良测量具有鲁棒性的自适应CKF。基于强跟踪滤波(strong tracking filter,STF)的理论框架和自适应CKF算法,构建强跟踪自适应CKF,该算法克服了标准强跟踪滤波的理论局限性,兼具STF与自适应CKF的优点,在系统存在模型不确定性及不良测量时具有良好的鲁棒性和滤波精确度。通过仿真验证了所建算法的有效性。     

Wind Velocity Sensorless Maximum Power Point Tracking Algorithm in Grid-connected Wind Energy Conversion System

This article proposes a wind velocity sensorless hill-climb searching algorithm in the grid-connected field-oriented controlled permanent magnet synchronous generator based wind energy conversion system. The proposed method improves the speed and accuracy of initial tracking. Pseudo-optimum points are used to track the exact optimum points using the power versus speed characteristics of a wind turbine for various wind velocities. This research emphasizes that the pseudo-optimum points using the power versus speed characteristics are always less than and much nearer to the exact optimum points. The small-step forward tracking after pseudo-optimum points improves the accuracy of this algorithm. Moreover, this method improves the tracking speed for various wind velocities. The system used for this research consists of a back-to-back sinusoidal pulse-width modulation converter between the generator and the grid. The performance of the proposed method is analyzed mathematically and verified by simulation using MATLAB/Simulink software (The Mathworks, Natick; Massachusetts, USA).     

基于目标跟踪的区域入侵检测方法研究

智能视频监控在人们生产生活中的应用越来越广,区域入侵检测是智能视频监控的一个重要方面。主要研究了基于目标跟踪的区域入侵检测方法。首先采用混合高斯模型进行运动物体检测,在实际验证中获得了较好的前景检测效果;进一步采用了Mean-Shift算法对从前景检测图像中提取出的运动物体进行实时跟踪,能够实时标记出运动物体并绘制出运动轨迹。采用OpenCV的运动物体跟踪算法框架,在Qt软件开发平台上进行了算法验证,获得了较好的实验效果,充分验证了算法的跟踪效果和实时性,具有实际工程指导价值。     

基于雷达信噪比模型的目标跟踪算法研究

为了解决目前雷达目标跟踪算法中存在的准确性低、稳定性差的问题,提出了一种基于雷达信噪比模型的雷达目标跟踪算法,该算法对雷达信噪比进行建模,评估测试过程中固定不变的参数数值,并对每一时刻的变量进行预估计算处理,从而获取目标运动轨迹和雷达跟踪强度。仿真结果表明,基于雷达信噪比模型的目标跟踪算法,确实可以及时有效的获取目标的运动轨迹和雷达跟踪强度,提高雷达目标跟踪系统的准确性和稳定性。