基于快速DBSCAN聚类的铁路异物侵限检测算法

在铁路建设及运营阶段,侵入限界的异物对既有线的安全行车构成极大威胁。将复杂的空间侵限检测转换为简单的平面内异物检测,研究利用二维激光测距传感器构建三维幕墙的侵限检测方法。在异物检测过程中,正常通过的列车不能认作异物,必须设计合理的算法检测并予以剔除。由于列车主要体现在轨道平面的检测结果中,重点研究通过分析扫描点云的分布特征对轨道平面内侵限物体进行分类的算法,提出了利用测量序列极值点作为核心对象的快速DBSCAN(density-based spatial clustering of applications with noise)聚类方法,并利用点簇的运动及分布特征判断是否为正常通过列车。现场试验表明,该方法能够有效区分侵限异物和正常通过的列车。

Intrusion detection algorithm for railway clearance with rapid DBSCAN clustering

In the stage of railway construction and operation, the objects intruding railway clearance greatly threaten the safety of running train on existing lines. In this paper, several 2D laser scanners are used to construct 3D laser curtains to detect intrusion in 3D space. Thus complex 3D intrusion detection is converted to simple inspection in plane. Besides intruding objects, the system must recognize and eliminate the effect of normal passing trains. A rapid DBSCAN(density-based spatial clustering of applications with noise) clustering algorithm is proposed, in which the extremum points of scan sequence are used as core objects of clustering. The movement and distribution characters of the clusters are used to judge if the cluster is a train or not. Field experiments show that the proposed method is an effective way to classify the intruding objects and passing train.