随着高速铁路6C检测监测系统技术规范的提出,基于视觉技术的非接触式接触网检测方法越来越受到业内研究人员的重视。为提高接触线上激光斑点跟踪定位的实时性以及测量值的准确性,提出了一种基于均值漂移和粒子滤波算法的接触网几何参数检测的新方法。首先,基于灰度颜色直方图特征分布和接触网"之"字形架构建立光斑目标模型;其次,利用聚类方法对粒子进行聚类,以聚类中心为起点运用均值漂移算法进行迭代计算,对迭代计算的结果利用粒子滤波算法得到光斑目标的图像坐标;然后,将激光斑点在图像坐标系下的坐标进行空间变换,得出接触线的几何参数导高和拉出值在世界坐标系下的测量值。最后,结合检测车在某供电段测试区的实际运行数据,验证了该方法的实时性和准确性。
With the publishing of 6C technology specification for high-speed rail detection monitoring system, the non-contact detection methods based on visual technologies for catenary system have received attention from researchers in the industry. To improve the real-time performance of tracking and positioning of laser spot on the contact line and the accuracy of the measured values, a new detection method for geometrical parameters of catenary system based on the mean shift and particle filter algorithm was proposed. Firstly, the model of laser spot target was established according to the feature distribution of gray color histogram and "Zigzag" structure of the catenary system. Secondly, the particles were clustered using the clustering method while iterative calcu- lation was conducted under the mean shift algorithm with cluster center as a starting point. The image coordi- nates of spot target were located by applying the particle filter algorithm for the iterative calculation results. Then, space transformation was conducted for the coordinates of laser spots in the image coordinate system to obtain the conductor height and stagger measurement values of catenary geometrical parameters in the world- coordinate system. Finally, the real-time performance and accuracy of the proposed method was verified based on the real operation data of detection vehicle in the test segment of a traction power supply section.