GML INspector™ is a desktop application for visualizing GML and CityGML data. City Geography Markup Language (CityGML) is becoming a de-facto standard for representation and exchange of 3D city models. GML INspector is designed to provide advanced capabilities for the examination and quality assurance of GML and CityGML data, and to support a variety of visualization modes, including both exterior and interior viewing. Unfortunately, the sheer volume and complexity of data contained in an average CityGML model makes effective visualisation of geometric, semantic, and topological relationships within a model very difficult.
Due to the complex interaction of the geometric, topological, and semantic properties in CityGML data, most currently-existing CityGML viewers focus almost exclusively on the visualization of appearance. Galdos developed GML INspector by leveraging the power of the Graphics Processing Unit (GPU) to go far beyond just appearance.
The modern GPU is not only a sophisticated visualization tool, but also a highly programmable and massively parallel general purpose computing unit. The ability to effectively utilize GPU adds a new dimension to interactive visualization and analysis of GIS data. GML INspector combines an effective user interface with several custom GPU rendering pipelines to provide a tool for the exploration of geometric, semantic, and topological relationships in a CityGML model.
GML INspector delivers some significant benefits for viewing CityGML data:
- Has a CPU to GPU pipeline that efficiently maps (tessellates) GML geometry elements to GPU primitives.
- Solves the problem of selection of nested screen objects by implementing a multi-pass GPU volume renderer coupled to geometry selection mechanisms.
- Has an embedded Python language interpreter that allows the user to access the data model directly and write sophisticated programmatic queries.
- Improves visual perception of geometry in indoor scenes by incorporating a specially-designed custom screen space ambient occlusion (SSAO) GPU renderer.
- Uses a WFS (Web Feature Service) client to fetch GML or CityGML through a WFS interface.
- Fast, efficient
- Constructs a very small memory footprint even for large CityGML files (e.g. 50Mb CityGMl file requires approximately 3Mb CPU RAM and 5Mb GPU RAM)
- Supports navigation of semantic features
- Written in C++
- Desktop installation – requires Windows XP and up
- Nvidia series 7 GeForce card
- Minimum 256 Mb RAM and up