Extending MonetDB with support for 3D Point Cloud and Voxels

MonetDB Solutions has joined a new project titled “Big Data Analytics in the Geo-Spatial Domain” in the framework of the Strategic Partnership Programme of the Netherlands eScience Center.

The project is proposed by Martin Kersten (CEO and Co-founder of MonetDB Solutions, and Research Fellow at CWI), together with Henk Scholten (SPINlab), Sisi Zlatanova (TU Delft) and Milena Ivanova (NLeSC).

As part of the project, we will help develop a pervasive and sound software infrastructures for Big Data Analytics, which is a key component in keeping abreast in the highly competitive eScience field. Many scientific domains need to manage and analyze big data collections with rich spatial and geo-spatial features. Among them large 3-dimensional point clouds collected by modern laser instruments gain increasing popularity in urban modeling, earth sciences, and digital object preservation of history and art.

During the last decade many DBMSs have been successfully extended with support for spatial and geo-spatial applications. However, the specific characteristics of point clouds and their sheer volume are not managed well by the traditional spatial features. New spatial data types and features are under development but the area is in its infancy and offers limited capabilities.

In this project MonetDB Solutions, under the lead of our new Senior Software Engineer Foteini Alvanaki, is going to extend the column-store analytical database MonetDB with storage and indexing techniques for 3D point clouds and 3D volumetric pixels (voxels) with properties, and provide functional primitives for building complex spatial analysis pipelines. The project will use a high-resolution urban modeling of the port of Rotterdam as a proof-of-concept application.