www.nathanbower.com

DTV - Introduction


DTV (Digital Terrain Visualisation) is the program that I chose to write for my third year project. I proposed the idea myself as I see a genuine need for a system such as the one I have developed, and the field interested me.

Java and Java3D were used to implement the system, so it is possible to run the application over the internet, or on any operating system that SUN has provided Java3D support for.


Concept


The general outline of the project was to create a system that displays a textured 3D model of parts of the UK, and allow the user to interact with them at real time. Essentially to take a map and make it into a 3d landscape that a user can fly about in.

In this way it will be extremely easy for someone to "see" what an area looks like, enabling planning, cognitive recognition etc. I realised that most people simply do not understand what contour lines are, and very few could accuratly draw the rise and fall of the land given only a contour map. My system abstracts this need, and allows the user to concerntrate on other aspects.


Development


The project timeline gave me about six months worth of development time, some of which was taken up with other tasks (revising for exams for example). In total I recon that I have put well over 500 hours into this project. Probably more like 750. The end result is well worth the "all night" programming sessions!

I realised that the program was something special when my freinds started popping by just to see how things were getting on!


Assessment


The only feedback that I have on the merit of the program was a demonstration, marked by a Doctor and Professor at Manchester university. My tutor commented that "very few" people ever acieve the ~80% boundary.

Mark Awarded = 78%

Context was clear?         [reasonably]
Context was Appropriate?   [definitly]
All Working?               [definitly]
A lot of Work?             [definitly]
High Technical Difficulty? [definitly]
(note - the catagories that can be choosen are definitly NOT, not really, adequatley, reasonable, and definitly)

Update: Final mark awareded for the project after the documentation had been marked 73%. Perhaps I shouldnt have done the documentation on a two night no sleep marathon?! Still well into 1st category though.


Continuation


I have demonstrated the project to several people in the GIS field at the national RS (remote sensing) society annual meeting in the DTI building London. Quite a few people were interested, only some were dissatisfied, and that seemed to be because my program put their companies 3d effort in a bad light.

The program has been used to visualise the Landsat datasets that MIMAS work with - and screenshots from these visualisations have been used on posters to promote the data service.

I have been working on a second edition of the program in my spare time. I have re-written the user interface, the DEM loader, the DEM interpreters, and have started work on J3D behaviors to dynamically load tiles as required, at an automatically selected level of detail. I have an aim to make this all work over the web at some point - but most computers will have a job using it because of the 3D and internet connection requirements, so I may make also build a cut-down version using a couple of square KM or so when I find the time.