论文部分内容阅读
从高分辨率影像中提取既有铁路中线,目前鲜有研究。铁路内部边缘丰富且具有良好的方向性,同时大量与线路近同向的铁轨阴影对中线提取造成干扰,使中线无法精确居中。提出一种两阶段的铁路中线提取方法,首先基于置信度方向直方图计算铁路中线的初始方向;然后根据轨间距与初始方向去除粗差点,以最小二乘法拟合铁路中线两侧最邻近铁轨线方程,通过近邻铁轨线方程内插实现铁路中线的精确定位。选取LiDAR影像中1 000m双轨高铁进行实验,结果表明:采用该方法能够准确、快速地提取实验区中的铁路中线。
There are few studies on extracting the existing railway centerline from high-resolution images. The interior of the railway is rich and has good directionality. Meanwhile, a large number of rail shadows near the same line interfere with the extraction of the center line, so that the center line can not be precisely centered. A two-stage railway center line extraction method is proposed. Firstly, the initial direction of the railway centerline is calculated based on the histogram of the confidence level. Then the rough points are removed according to the distance between the track and the initial direction. The least squares method is used to fit the nearest railway line Equation, the precise location of the railway centerline can be achieved by the interpolation of the nearest railway line equation. The experiment of 1 000 m dual-track high-speed rail in LiDAR image was selected. The results show that the method can extract the railway center line in the experimental area accurately and quickly.