Classification of airborne 3D point clouds regarding separation of vegetation in complex environments

Abstract

Classification of outdoor point clouds is an intensely studied topic, particularly with respect to the separation of vegetation from the terrain and manmade structures. In the presence of many overhanging and vertical structures, the (relative) height is no longer a reliable criterion for such a separation. An alternative would be to apply supervised classification; however, thousands of examples are typically required for appropriate training. In this paper, an unsupervised and rotation-invariant method is presented and evaluated for three datasets with very different characteristics. The method allows us to detect planar patches by filtering and clustering so-called superpoints, whereby the well-known but suitably modified random sampling and consensus (RANSAC) approach plays a key role for plane estimation in outlier-rich data. The performance of our method is compared to that produced by supervised classifiers common for remote sensing settings: random forest as learner and feature sets for point cloud processing, like covariance-based features or point descriptors. It is shown that for point clouds resulting from airborne laser scans, the detection accuracy of the proposed method is over 96% and, as such, higher than that of standard supervised classification approaches. Because of artifacts caused by interpolation during 3D stereo matching, the overall accuracy was lower for photogrammetric point clouds (74–77%). However, using additional salient features, such as the normalized green–red difference index, the results became more accurate and less dependent on the data source.

Ground Filtering of Co-Registered Mobile and Stationary Laser Scans by Using Superpoints in Ransac Planes

“…If we balance the resolution for evaluation and compare the result only for the more challenging segment 1, the numerical improvement is clearer (Tables 3 and 4). All tests for Tables 1 to 4 were carried out with the superpoint resolution 0.1 m, which turned out to be a quite good choice for this dataset and has the same order of magnitude as for the photogrammetric dataset in (Bulatov et al, 2021). Experiments with larger resolutions, up to about 1 m, perform increasingly better on recognizing correctly more street points, especially with sparse point cloud density in areas away from the scan.…”

“…The advantage of SiRP is that it can handle both purely 3D structures of data and scarcity, or even absence, of training examples. In (Bulatov et al, 2021), this method was applied to both an actively-sensed airborne point cloud and a result of a photogrammetric reconstruction from a UAV-borne sequence of high-resolution images. The UAV flight around the Gubbio wall (Italy) at a moderate altitude ensured a sufficient overlap of images.…”

Filtering Green Vegetation Out from Colored Point Clouds of Rocky Terrains Based on Various Vegetation Indices: Comparison of Simple Statistical Methods, Support Vector Machine, and Neural Network

“…Various geometric and structural filters have been implemented in commercial or non-commercial software for use with LiDAR point clouds [8] or point clouds generated photogrammetrically [9]. New filtering algorithms or procedures have also been developed [10][11][12][13][14], some of which have been successfully applied to rock masses [15,16]. However, the successful use of algorithms with the geometric principle for rocky terrains is difficult due to the ruggedness of the formations.…”