www.storm.net.nz

[ / ] [ Metlstorm ] [ Projects ] [ ice.storm.net.nz ] [ \m/ ]

Projects
[ Security: SSH 'Jack Hai2IVR MAFL-Load Firewire, DMA & Windows Asterisk Remote Root Metl-o-UnNetCrypt ] [ Wireless: Metl Kismet GPS Plotter - Google Earth Edition Metl War Tri Pod Metl Kismet Client Metlstorms Kismet GPS Plott0r Metl Helix Wireless Grapher Metl Network Recon Visualizer ] [ Home: Rotoseat Noise Weblstorm Viewtron CharGrill ] [ Abandonware: Obscured By Clouds ]

Metlstorms Kismet GPS Plott0r
MKGP was my first foray in to geospatial visualization of wardriving data. I'd got plenty of georeferenced Kismet logs from wardriving with the Helix boys, but just looking at XML files doesn't give me wood.

I started with the map plotter that is in the Kismet distribution, but soon discovered that it was based on GPSDrive, which is pretty fundamentally broken. I of course, didn't know this at the time, but in the process of proving so to myself I ended up writing a raster mapping engine in python. I toyed with a couple of realtime moving map display projects, which you can see elsewhere, but ended up using the code to write MKGP, a map vizualiser for Kismet.

One of the particular requirements is that it had to work in New Zealand, that is, with our crazy domestic map reference system. I didn't understand yet quite the horror this meant (I learnt from my mistakes for the 3d map viz engine for MWTP, however.)

Despite being new to this whole GIS thing, MKGP is quite usable in NZ. It auto pulls maps from the government map server and has a python based query language for filtering your data. It has all manner of configurable display knobs, for rendering approximated range circles, packet-to-source lines, and all that sort of guff.
Last Update: 2006-12-18 21:38:11
State: Usable
Distribution: Not yet public
Tags: Wireless
Images:
An early plot (note lack of alpha blending on the blobs) showing the coverage of local wireless ISP Walker Wireless (now <a href="http://www.woosh.co.nz">Woosh</a>, and mostly out of the 802.11 game, in favour of WCDMA). Because I know the location of their transmitters, it's a good way to get some idea for the (in)accuracy of my AP-location code. You'll note that it's not very good. I have some ideas to improve it, but fundamentally it's very difficult to do location finding in an urban environment with omnidirectional kit. Here MKGP renders a query for all AP's who observed range is less than 500m in downtown Auckland (yes, there's that many APs in Auckland city. So far over the last couple of years I've seen over 3000 unique networks). You can see range circles, coloured by channel, and some lines plotting where we recieved a packet to where the AP is predicted to be. The blue line is the GPS track, and the black blobs are APs. I've also toyed with plotting on orthophotos, but getting hold of good quality, well-corrected arial photos isn't easy when your budget is $0. I referenced these by hand, so the plot-track is a bit wonky.