Re: Longest distance between camera and PC?

["accidental plumber" <aplayerinla@xxxxxxxxx>]

You have my sympathy.  I'm not going to lecture the need for deterrence
and protection again.  But my personal experience is that - things
escalate.  When I installed PIR floodlights they do it right after
school.  It was pure harmless prank for now but I had to clean up.  At
least think about what worse they could do to your insecure(?) outlying
buildings on sand(?), if they fail to steal?

(PeteCresswell) wrote:
>
> Thanks for the detailed and informative reply.   You just told me more than I
> could ever hope to glean in a week's worth of Googling.

The chance of meeting an accidental security expert on vacation at home
is not high.  What I do is froogle instead of google.  Get some junks
with good deals at eBay to learn something before commit to expensive
systems (if ever).  There's the option of returning the junk by paying
a good portion of it, or safe the junk for later days.

>
> As far as requirements go, the primary requirement is to allow a total
> dilettante (myself) play around with TV/motion detection implemented via a PC.
> The secondary requirement has nothing to do with protection.  Vengeance is more
> like it.

Vengeance, exactly what I wanted.  Mess up my front lawn while I took a
nap was a challenge.  Though it took less than 5 min to clean up; they
dont' have much time in daytime.

But then after I experience with night vision, I began to realize the
difficulty relative to deterrence.  I still have cameras, but to
achieve my original aim it would cost me a lot more with doubtful
results.  OK, no lecture again.  Just one thing, software motion
detection doesn't go well with night vision!

>
> A friend has a small windsurfing retail shop on the beach on the edges of a
> multi-story hotel's parking lot.   He's got three small buildings: two
> side-by-side and one facing the two.   Next to the two side-by-side he has
> rental racks where customers store windsurfers.

Wireless is a real option here.  The buildings aren't big.  You can
keep mostly line-of-sight between outdoor cameras and the receiver.
There aren't much interference at the beach at night.  There are too
many ways to make cameras useless other than jamming them.  I assume
the thieves aren't sophisticated.   Even crappy tiny cameras costing
$30 will have flawless reception if you place them well - short line of
sight.  Wi-fi has a much better range - digital my friend!  Your router
can be put in a more secure place further away from the scene.  And you
can monitor them at your home rather than at your friend's business.
Video motion detection is rather tolerant to flickers of wireless cams.
 It's too fast for it to detect.  And the sensitivity can be adjusted
anyway.

>
> The vengeance part comes in if/when we can identify whoever has been stealing
> windsurfers from the racks (bolt cutters... and so forth...).
>
> These are felony thefts and given clear photos of the perps and a shot of the
> vehicle - preferably with legible license# - I'd hope the police would regard
> such a case as low-hanging fruit and go after it.
>

Well that's draw of the luck.  It depends on whether the perps and
their cars are whether you want them to be.  You know, as you say, even
banks aren't doing it well.  Whatever technology you use, it's like a
video shoot by a 1/3 to 1/4 camcorder at outdoor light at some
distance.  You try to do some video and see.  Banks are well lit.

>
> So far, the scenario doesn't seem all that unrealistic.
>
> To wit:
> ------------------------------------------------------------------------------
> - The rack area is lighted and can be lighted even more, no problem.

My camcorder at night at outdoor lights is pretty poor.  So I brought a
night vision, different kind of problems and not any better.  (Though
it may be better if you throw in more money.)  The low contrast image
couldn't even trigger the motion detection via video processing.  I
found out that I need bright light to turn on to generate enough
'motion' for the software.  This is close to floodlight so I just as
well move the PIR floodlight near the camera and install more PIR
floodlights.  As I have mentioned elsewhere.  No more morons walking
their dogs after 10 pm near my house, joggers returning late would
cross the road first and proceed rather than passing in front of my
house.

>
> - All three buildings have AC power and are weather tight: no problem
>   having a PC in them.  In fact one building has had a weather sensor PC
>   in it for some years now with no problems.

Software solutions for a 'businessman' isn't suitable, unless he's geek
also.  The majority of small business use time lapse VCR.  So whoever
minimum wage employee had to do is to know how to change and rotate the
VCR tape at regular intervals.  Ordinary PC's running 24/7 with video
coder on is just too hot.  In contrast, my cheap cmos cameras are
possibly running constantly for years without problem.

>
> - The distance from the facing building to the rack area is less than 25 feet.

Wireless?

>
> - Due the visibility and the size of the things being stolen, we'd be hoping
>   that somebody would back their vehicle into the area closest to the racks.
> This make sense in that they might feel more comfortable doing the deed mostly
>   screened from view by the facing building - as opposed to carrying the boards
>   across the hotel's parking lot at night.
> --------------------------------------------------------------------------------

Almost all expensive security cameras are designed to deter as well,
large and intimidating.  For night vision, I have seen 36 high power IR
LED on a plate and 72 LED's as the light source.  They ain't small and
invisible.  So I would be thinking of small CCD's with good lighting.

>
>
> What I'm coming up with so far is:
> --------------------------------------------------------------------------------
> 1) Multiple cameras.  One camera on the rack area.   Two or three
>    others to cover possible parking areas and access/egress to/from the
>    rack area.
>
>    The main problem would seem tb selection of instruments that will give
>    the needed definition.
>
>    I've seen too many clips on the evening news from retail store security
>    cameras that had a full frontal shot of somebody holding up the store,
>    but in which the person is still totally unrecognizable.  I *think*
>    I've even seen the same scenario from bank security cameras - which one
>    would hope are at some professional level of quality.

Unless for expensive custom technology, who would be bulky anyway, the
technology is limited to mass consumer technology.  Most CCD's are the
same used in camcorders.  Even for camcorder quality, DVD quality at
720x480, the chance that you see clearly the face of a person is low,
as in your home movies.  If you zoom in to get a clear head shot or
license plate, the coverage area is small, and vice versa.  There must
be exceptions, possibly in the HD area, but rarely any camera has
resolution higher than DVD quality.  Unless for wifi cam, which can be
had for $100 anyway, the interface is possibly limited to NTSC.  There
can be many lines of resolution, >1000 for big screen TVs, but the
number of horizontal scan lines is fixed, since the dawn of TV.

So you made me doubt if video is ever the solution.  You can have a
very cheap 3 Mpix camera to take much better pictures.  Now the
resolution goes to 10 Mpix?  You can have a printed picture the size of
a wall without grains?  Some cam already sacrifice high resolution to
give a better picture at low lights.  So for high resolution I will be
next time looking at cams that would take high resolution snapshots, or
something like 10 fps.

>
> 2) A husky-enough PC/interface card to support the concurrent frame rates.
>    I'm thinking full motion (30 fps?) if only because it would make a better
>    impression on whoever the local police send to investigate.

You can froogle around and see.  But video coders (mp2/mp4) are
designed to be much more complex than decorders.  So the DVD can run
long time and keep cool, but the video cards can't.  Mp4 should be
"hotter" than mp2.  Somehow if your signal comes in digitally, the bit
rate are still pretty hot; cable modems have big cases to dissipate the
heat.  The bottleneck is the PCI (express) bus on the PC, which could
or could not stand the 24/7 operation.  Same for the hard disk - your
requirement easily surpass that of a company with IT guys to maintain
the file servers.

>
> 3) Hard wiring.   No problem going from building-to-building if needed.

Depends what type of video stars you have.  You can jam the wifi, the
RF, or you can throw stones an use a big cutter.

>
> 4) A timer, of course... It would be futile to have this thing running during
>    business hours.  I'm assuming that this is a standard option in virtually
>    all camera control software.
>
> 5) Motion sensing and a couple of 500-gig hard drives.   Overkill, for sure,
>    but the reality is that this thing will probably be forgotten about for
>    months and months on end, only to be reviewed if/when another theft is
>    committed - and I suspect that we will get our share of seagulls walking
>    by, small animals sniffing around, and customers dropping by when the
>    shop is closed.
> ---------------------------------------------------------------------------------
> --

Very CSI or Las Vegas.  1G is about 4 hours for mp4.  So you have 2K
hours, 10 hr/day gives you 200 days?  Or 50 days for mp2, DVD quality.
With 5 cams you left with only 10 days.  Though with motion or lower
frame rate you have plenty days more.

Firstly I doubt if anyone walking pass your store a couple of months
before will prove anything.  Find the associates who get caught?

I doubt if anything not designed to run continuously, like a PC, will
stand that sort of torture.  A security DVR is under $1000?  $700?  But
the total frame rate is probably 30 fps and the storage is limited.  I
would think 24 hrs of storage is minimum.  You dig up the video when
you discover the break in the next day.  A few days is good and a week
or more is best allowing you go on vacation without losing video.

Realtime software motion detection is probably too hot.  You need to
decode the 'digital' feed, average some frames and subtract some
frames, then count the pixels.  In contrast, it's trivial doing it
after digitizing the analog signal with simple hardware.

And PIR is pretty fail safe.  It does not detect heat, but change in
light, which a person is like a torch.  That gives you some rest for
your hardware too.  Say the PIR's tell the recorder which camera should
be recorded.

I'm still a little disappointed.  With all the PIR and trigged
lighting, I can't catch even innocent people walking pass on cam.
Should I install an x10 dimmers?