基于扫描线模型的机载激光点云滤波算法

Filtering Algorithm for Airborne Laser Point Cloud Based on Scanline Model

文章针对现有滤波方法在阶跃地形处滤波的不足,提出一种基于扫描线模型的机载点云滤波新算法。首先对数据X坐标等间距沿同一方向扫描点云,生成扫描线组,构建等边三角形模型,滤除每条线上的地物点;继而对滤波后的地面点以Y坐标等间距构建矩形格网,运用最小二乘法进行曲面拟合;根据拟合高程与实际高程的比较结果判断属性,滤除地物点。通过实验证明文章的滤波方法可有效达到预期目标,滤除建筑物和地物点,适用性好。

In this paper, a new airborne point cloud filtering algorithm based on scanning line model is proposed to overcome the shortcomings of the existing filtering methods in step terrain. Firstly, the point cloud is scanned in the same direction according to X-coordinate, so that scanning line group is generated, equilateral triangle model is constructed, and object points on each line are filtered out. Then, rectangular grid is constructed with Y-coordinate equidistant distance to filtered ground point, and the least square method is adopted to fit the surface. According to the comparison between the fitting elevation and the actual elevation, the attributes are judged, and the ground feature points are filtered. The experiment proves that the filtering method in this paper can effectively achieve the desired goal, filter out buildings and ground features, and has good applicability.