Dellé Odeleye wrote:
> I'm not aware of any challenges being made to the accuracy of their
> methodology
I don't have any problem with their *methodology*. But I've seen
first-hand their performance fail in some situations, and succeed in others.
> I think there's another company as well, using a similar technique.
There are several, Footfall being the most prominent, and Qinetiq being
one of the most innovative. Dr. Jia Hong Yin has his own software which
has been tested with success on Oxford Street. AIUI, Springboard
re-badge the Qinetiq software (for their cctv tools - other software for
their IR and lasers) and add their own front-end. There's also InfoDev
in Canada, though I haven't seen their kit first-hand. The Lase Peco kit
is about to be trialled in Central London, and looks promising on paper
(as they all do, of course).
There are trials of a radar that is normally used for classifying
vehicle types and speed: it works well for counting bicycles, but to
calibrate it for pedestrians is to operate it outside the design spec,
really. Nevertheless, it may offer an interesting alternative, as it's a
compact unit requiring only a DC power supply, so could easily be
mounted on traffic light, for example.
One should bear in mind that reliability in shopping centres, airports
and the like is no guide at all to the reliability on-street in general
use. Installations at shopping centres typically offer high-quality
lighting, higher control over the monitored environment generally, mains
power, secure storage facilities for a pc, and often a dedicated phone
line. Such luxuries are rarely available to a local authority at typical
street locations. (well, the mains is, of course, but then you get into
the entertaining issue of discussing with the lighting engineers the use
of their columns).
Andrew Smith
--
Director, London Analytics
T: [+44] (0)20 7627 8924
M: [+44] (0)791 046 0601
W: http://www.LondonAnalytics.info/
|