Humber in a Box is the work of four Computer Science post-graduate students from the University of Hull, who worked on the application as part of a project for their Masters course. This process began in October 2014 and finished in Spring 2015 – debuting at the Hull Science Festival.
Thank you to John van Rij, Leo Abbas, Danny Quarmby and Benjamin Allison.
Collection of images from the development of Humber in a Box
The application itself uses the UNITY-3D gaming engine, into which the hydraulic code from CAESAR-Lisflood was merged.The model of the Humber is built from a Digital Elevation Model (DEM) at 500 m pixel resolution (or ‘nodes’ to the games developers), which was built from data provided by the Environment Agency (EA) and the Association of British Ports (ABP) for the Dynamic Humber Project (DHP). Into this DEM, the heights of all the sea defences around the Humber are incorporated.
The tides that are seen in the model are driven by recorded tidal stage data sampled from the ABP gauge at Spurn Point in 2010. This is used to raise and lower the water level at the North Sea end of the model, with the internal model physics simulating the flows around the Estuary.
Images from a prototype testing day during development
The participant views this model in beautifully rendered 3D graphics, positioned on top of a box in the middle of a virtual museum room. This is viewed via an Oculus Rift Developer Kit 2 Virtual Reality (VR) headset, immersing them fully into this room. It is hard to describe without trying it, but it really like you are there! They can explore the museum, browsing the images of the Humber in ornate frames, or interact with the model.
Movement is controlled using an XBox controller, and the participant can change their view using either the controller, or moving their heads – the Oculus is fully motioned track and responds to any head movements.
BBC Look North Weatherman Paul Hudson tries an early version of Humber in a Box
One of the challenges to this project was computational power. Hydraulic models are computationally expensive and we use powerful 6-core i7 machines. In addition, the machine also needs to produce the 3D graphics at a framerate of at least 75fps otherwise their will be a lag with the motion tracker – this will make participants feel dizzy and ill! The team did a great job getting the balance right for this.