长江中下游河道岸滩低空机载LiDAR点云地形滤波算法

探索新兴的低空机载Li DAR技术,解决长江中下游植被高覆盖、高遮挡区岸滩地形测绘难题。提出一种针对多层次、高密度植被覆盖区的低空机载Li DAR点云植被滤波算法,该算法通过多回波分析对点云进行粗滤波后,采用数学形态学运算获取地面种子点,对地形局部进行趋势面拟合,再通过随机采样一致性检测,剔除植被点,保留地面点,从而获取测区的数字高程模型(DEM)。典型测区试验表明,该滤波算法能解决长江中下游河道岸滩地区地形起伏较大、植被高覆盖区域植被点云智能化剥离难题。根据测区实际情况,设计针对性的滤波算法,即使是植被与地面点云高度混淆、激光穿透率低于15%的复杂情形,仍能有效分离出地面和非地面点。

Filtering algorithm of the low-attitude airborne LiDAR point clouds for topographic survey of the middle-lower Yangtze River riparian zone

This paper firstly explores and applies the newly low-altitude airborne LiDAR technology to solve the mapping problem of the middle-lower Yangtze River Riparian zone covered with multi-layer and high-density vegetation. A vegetation filtering algorithm is proposed for the low-attitude airborne LiDAR point clouds collected in the complex area with multi level and dense vegetation coverage. After the coarse filtering of point clouds by multi-echo analysis, this algorithm extracts the ground seed points by morphological calculation, fits the trend surface of the local terrain, then eliminates the vegetation points and preserves the ground points using the Random Sample Consensus, thereby acquiring the DEM of measured area. The experimental results show that the proposed algorithm can intelligently remove the vegetation point clouds in the middle-lower Yangtze River riparian zone, where topographic relief is large and vegetation cover is dense. This research also shows that by designing a special filtering approach, it is possible to classify laser points into terrain and vegetation automatically even for thoroughly mixed vegetation and terrain points with low penetration rate of below 15%.