Mining lidar data with spatial clustering algorithms

Clustering algorithms have been an important area of research in the domain of computer science for data mining of patterns in various kinds of data. This process can identify major patterns or trends without any supervisory information such as data labels. Broadly specified, it divides a set of objects into clusters each of which is a representative of a meaningful sub-population. In this article, we first carry out an assessment of available categories of clustering techniques and find that hierarchical- and density-based algorithms are apt for clustering light detection and ranging (lidar) data. Then, we adapt and examine the effect of two algorithms, namely density-based spatial clustering of applications with noise (DBSCAN) and ordering of points to identify the clustering structure (based on perimeter of triangles) (OPTICS (BOPT)) found in the literature in the area of knowledge discovery in databases, on lidar data. The performances of the algorithms are evaluated with respect to execution time and comparison of clustering outputs with respect to a manually classified data set of . DBSCAN performs better in both respects. The efficacy of DBSCAN is also demonstrated for detecting clusters of complex shapes for two different data sets, each of areal dimension of .