STR-octree indexing method for processing LiDAR data

There are increasing attentions on the emergence of large-scale spatial data generated by various types of sensors in geospatial and computer science studies. However, processing the large-scale spatial data is a challenging issue because of its speed, size, and non-uniform distribution. The traditional methods are not able to load, access, and process the data directly in the big data environment due to hardware limitations. In this paper, we propose a novel indexing approach, called STR (Sort Tile Recursive)-octree, to process large-scale spatial data generated by LiDAR (Light Detection And Ranging) sensors. We also propose a high-performance processing architecture that can be applied to a variety of parallel framework. The proposed approach was evaluated with airborne LiDAR datasets in various point distributions, and the results show effectiveness of our approach especially in non-uniformly distributed datasets. The approach can be generalized to be utilized in other spatial data sets including higher-dimensional cases.