After reading the Feature Proposal at
https://wiki.openstreetmap.org/wiki/Proposed_features/Extended_tags_for_Key:Surveillance
it
seems that I want this proposal to open up the possibility of mapping the
broadest possible spectrum of surveillance might indicate changing the
proposal Key to the broader term of 'Sensor'. 'Surveillance' describes a
very specific negative connotation. Also, even in the case of video, there
can be audio present, which indicates that at any location there might be a
wide variety of sensors at a particular location. For example, some of our
local weather stations have webcam feeds so one can actually see what the
weather is like, along with the actual 'data' feeds. A video camera is just
one sort of sensor, and reading the FP it seems you accommodated this with
the Surveillance item specific tags. For that set of tags, there are at
least several ontologies and standards out there that could provide you
with exhaustive sets of concepts you could then extract from. From the W3C
at http://www.w3.org/2005/Incubator/ssn/wiki/Report_Work_on_the_SSN_ontology :
,ie. which perspective are you using?
The ontology can be used for a focus on any (or a combination) of a number
of perspectives:
- A sensor perspective, with a focus on what senses, how it senses, and
what is sensed;
- A data or observation perspective, with a focus on observations and
related metadata;
- A system perspective, with a focus on systems of sensors; or,
- A feature and property perspective, with a focus on features,
properties of them, and what can sense those properties.
Let me know if you have things to say ;)
It seems that lumping things like guards and dogs in with sensors is
awkward at best. Would this then include Neighborhood Watch group members,
for instance? Even the word guard is problematic, how could someone tell a
grounds caretaker from an armed guard, or someone in a business suit that
could be a US Marshal or a well dressed receptionist. This would seem
better handled separately in some sort of Law Enforcement ( official,
private, para, public ) tag set.
It would seem that word 'surveillance' rather than being the major key it
should instead be be relegated to the bottom of the hierarchy as one choice
of perhaps multiple functional intents, for instance ( forgive the ad hoc
syntax), sensor( type:video ( observation:trafficcontrol, observation:
trafficenforcement, observation:surveilance , observation:
energymanagement, observation:pigeonfeeding )) in the next decade sensors
(including video) are going to be ubiquitous, and will probably far
outnumber the Law Enforcement cctv (see
http://www.ict-sensei.org/index.php?option=com_frontpageItemid=48 for
example). Since the same sensor may feed multiple networks for each type of
observation you might specify which agency is getting that data, police,
corporate, EPA, etc. if known.
Nice, finally the open burglary map comes closer ;-)
First thought in my mind was the law of unintended consequences, where this
allows the Orwellian Overlords to progressively refine their efforts,
eventually using concealed cameras to tape unsuspecting miscreants in the
few remaining areas.
2. Isn't there one parameter missing to deduct the actual area covered by
a camera? I'd think you needed 3 values: direction the camera
points to (in 3d, e.g. azimuth and altitude) plus the field of view or
focal length. This is of course purely theoretical because the cameras
might be able to move and most mappers won't probably be able to add high
precision orientation data (usually you will have to
estimate these values).
While you address camera capabilities, what is probably more significant to
the public is the scene viewed by the camera. If you wanted to hairsplit
this aspect further for video, there are several active projects with
proposed ontologies for scene description and field of view, etc. that
could be the basis of more detailed tags.
http://www-sop.inria.fr/members/Francois.Bremond/topicsText/advisorProject.html
Again, you could subset these to make a doable tagset for ordinary mappers.
My guess is that the machine vision people already have established
something simple for outdoor areas ( for instance, autonomous vehicles use
outside cameras in a location if available for scene refinement.
These and similar ontologies, standards, metadata appear intimidating at
first, and very complex, but they are usually the results of collaborative
effort of hundreds of knowledgeable people over years, and serve as at
least a starting point, carve out the rest, and leave the best for your
purposes. One or a few individuals can rarely conceive of all the
overlapping properties and aspects of physical objects ( especially on a
worldwide basis, IMHO you can get a head start by not re-inventing the
wheel.
Michael Patrick
___
Tagging mailing list
Tagging@openstreetmap.org
