LiDAR data management pipeline; from spatial database population to web-application visualization

While the existence of very large and scalable Database Management Systems (DBMSs) is well recognized, it is the usage and extension of these technologies to managing spatial data that has seen increasing amounts of research work in recent years. A focused area of this research work involves the handling of very high resolution Light Detection and Ranging (LiDAR) data. While LiDAR has many real world applications, it is usually the purview of organizations interested in capturing and monitoring our environment where it has become pervasive. In many of these cases, it has now become the de facto minimum standard expected when a need to acquire very detailed 3D spatial data is required. However, significant challenges exist when working with these data sources, from data storage to feature extraction through to data segmentation all presenting challenges relating to the very large volumes of data that exist. In this paper, we present the complete LiDAR data pipeline as managed in our spatial database framework. This involves three distinct sections, populating the database, building a spatial hierarchy that describes the available data sources, and spatially segmenting data based on user requirements which generates a visualization of these data in a WebGL enabled web-application viewer. All work presented is in an experimental results context where we show how this approach is runtime efficient given the very large volumes of LiDAR data that are being managed.