Azul: A fast and efficient 3D city model viewer for macOS

Ken Arroyo Ohori

doi:10.1111/tgis.12673

Abstract

Abstract3D city models are an important research topic within geographic information, but there is still a lack of good tools to work with them in practice. In an attempt to alleviate this problem and to help with our own research, we have developed azul, a free and open‐source macOS 3D viewer that was especially engineered for the visualization of 3D city models. The aim of this article is, first of all, to describe the inner workings of azul as a complete methodology to efficiently visualize 3D city models, and which can be also applied to other hierarchically structured data models, which are becoming more prevalent in geographic information. In addition, the article has three ancillary goals: (a) to present other general‐purpose methods that are useful to efficiently process 3D city models (e.g., robust error‐correcting parsers and data models that can be used with multiple formats); (b) to describe technical issues and problematic aspects related to current 3D city model formats that are known by developers but are not properly documented in the scientific literature; and (c) to foster an open discussion about the best data structures and algorithms to process 3D city models in practice.

Highlights

More open 3D models of cities are becoming available,1 and these are being used in an increasing number of applications (Biljecki, Stoter, Ledoux, Zlatanova, & Çöltekin, 2015)
The aim of this article is, first of all, to describe the inner workings of azul as a complete methodology to efficiently visualize 3D city models, and which can be applied to other hierarchically structured data models, which are becoming more prevalent in geographic information
The article has three ancillary goals: (a) to present other general-purpose methods that are useful to efficiently process 3D city models; (b) to describe technical issues and problematic aspects related to current 3D city model formats that are known by developers but are not properly documented in the scientific literature; and (c) to foster an open discussion about the best data structures and algorithms to process 3D city models in practice

Summary

| INTRODUCTION

More open 3D models of cities are becoming available, and these are being used in an increasing number of applications (Biljecki, Stoter, Ledoux, Zlatanova, & Çöltekin, 2015). In addition to good Mac support, we had a few other important software design goals based on the main problems we have found when working with 3D city models: (a) support for all the 3D city model formats that we regularly use in our research; (b) fast loading and display of models; (c) sufficiently low memory use to smoothly display large areas at once; (d) robust methods that gracefully handle the typical errors present in most models; and (e) intuitive interaction that takes into account the particularities of 3D city models (e.g., a large horizontal and small vertical extent) and those of standard Mac hardware (e.g., multi-touch trackpads rather than mice with multiple buttons and scroll-wheels) These goals should be built on top of a mostly C++ base for future cross-platform compatibility.

| BACKGROUND

Findings

| CONCLUSIONS AND DISCUSSION