"Hiding" is the wrong word. How about "making cameras less
of an eyesore"? (I despise having *any* kit visible!)

I'm in the process of updating my "user tracking" system
to rely on vision (prototype was RF based but that leaves
open a potential attack vector). So, I need to install
a boatload of cameras. Most of them *inside* (home,
commercial establishments, etc.)

Just mounting them on walls is easy -- but, makes it look
like you're in a bank lobby!

I'd thought of down-facing cameras but then I'm limited
to tracking motion; "features" would tend to be less visible
(unless perhaps using a 360 degree prism? Likely a
computational nightmare!). Ideally, I'd like to be able to
recognize individual users instead of just "presence detect".

The other issue is that they have to be accessible for
service. The easiest way would be to allow service "from
the rear" -- another advantage for down-facing (if "the
rear" is *attic* space)

I've toyed with some small "spy" cameras with tiny apertures.
But, they tend to have a fair bit of bulk *behind* the lens.

I'll skip over illuminator requirements; worry about if that
will be an issue *later*!

I just can't see a portion of the solution space that would
address all of these issues (neglecting cost). Maybe
work strictly in Ir and use some sort of semiopaque lens
to disguise its presence? (I wonder how transparent a thin
layer of drywall compound would be to Ir?)

On Sun, 1 Aug 2021 13:42:15 -0700, Don Y <blockedofcourse@foo.invalid>
wrote:

>"Hiding" is the wrong word. How about "making cameras less
>of an eyesore"? (I despise having *any* kit visible!)
>
>I'm in the process of updating my "user tracking" system
>to rely on vision (prototype was RF based but that leaves
>open a potential attack vector). So, I need to install
>a boatload of cameras. Most of them *inside* (home,
>commercial establishments, etc.)
>
>Just mounting them on walls is easy -- but, makes it look
>like you're in a bank lobby!
>
>I'd thought of down-facing cameras but then I'm limited
>to tracking motion; "features" would tend to be less visible
>(unless perhaps using a 360 degree prism? Likely a
>computational nightmare!). Ideally, I'd like to be able to
>recognize individual users instead of just "presence detect".
>
>The other issue is that they have to be accessible for
>service. The easiest way would be to allow service "from
>the rear" -- another advantage for down-facing (if "the
>rear" is *attic* space)
>
>I've toyed with some small "spy" cameras with tiny apertures.
>But, they tend to have a fair bit of bulk *behind* the lens.
>
>I'll skip over illuminator requirements; worry about if that
>will be an issue *later*!
>
>I just can't see a portion of the solution space that would
>address all of these issues (neglecting cost). Maybe
>work strictly in Ir and use some sort of semiopaque lens
>to disguise its presence? (I wonder how transparent a thin
>layer of drywall compound would be to Ir?)

Have a look at some of the spy camera setups if you're interested in
that sort of thing- mounting in smoke detectors, in wall power
outlets, in coat hooks, in clock radios where there is natural space
behind the aperture. I'm sure paintings are another possibility.

Eg. https://amzn.to/3C6Hi15
--
Best regards,
Spehro Pefhany

On 8/1/2021 2:04 PM, Spehro Pefhany wrote:
> On Sun, 1 Aug 2021 13:42:15 -0700, Don Y <blockedofcourse@foo.invalid>
> wrote:
>
>> "Hiding" is the wrong word. How about "making cameras less
>> of an eyesore"? (I despise having *any* kit visible!)
>>
>> I'm in the process of updating my "user tracking" system
>> to rely on vision (prototype was RF based but that leaves
>> open a potential attack vector). So, I need to install
>> a boatload of cameras. Most of them *inside* (home,
>> commercial establishments, etc.)
>>
>> Just mounting them on walls is easy -- but, makes it look
>> like you're in a bank lobby!
>>
>> I'd thought of down-facing cameras but then I'm limited
>> to tracking motion; "features" would tend to be less visible
>> (unless perhaps using a 360 degree prism? Likely a
>> computational nightmare!). Ideally, I'd like to be able to
>> recognize individual users instead of just "presence detect".
>>
>> The other issue is that they have to be accessible for
>> service. The easiest way would be to allow service "from
>> the rear" -- another advantage for down-facing (if "the
>> rear" is *attic* space)
>>
>> I've toyed with some small "spy" cameras with tiny apertures.
>> But, they tend to have a fair bit of bulk *behind* the lens.
>>
>> I'll skip over illuminator requirements; worry about if that
>> will be an issue *later*!
>>
>> I just can't see a portion of the solution space that would
>> address all of these issues (neglecting cost). Maybe
>> work strictly in Ir and use some sort of semiopaque lens
>> to disguise its presence? (I wonder how transparent a thin
>> layer of drywall compound would be to Ir?)
>
> Have a look at some of the spy camera setups if you're interested in
> that sort of thing- mounting in smoke detectors, in wall power
> outlets, in coat hooks, in clock radios where there is natural space
> behind the aperture. I'm sure paintings are another possibility.

I've already done that in some novelty items. E.g., I have several
"worlds of wonder" animatronic toys scattered around the house that
act as speakerphones (so the eyes and mouths of Mickey/Goofy/Teddy/etc.
appear to react to the caller's words); relatviely easy to squeeze
a small CCD camera behind their "eyes", in their nose, etc.

But, I was hoping for a more "universally deployable" solution. I.e.,
"I want a camera *here* to give me coverage of *this* area from *this*
vantage point and a WoW 'toy' would look silly hanging on the wall!"
And, a room littered with such toys would be sort of creepy!

[My wife's grandmother collected antique, bisque dolls -- each about
2 feet tall. For display purposes, they'd be mounted on a stand as
if standing, fully costumed. Hands would be bound together as they're
accidental banging into each other (or nearby objects) would shatter
the porcelain. Imagine *hundreds* of the little critters lined up
in neat rows and columns. All facing the back door (used by family
members, only). You'd walk in and get the heebie-jeebies with all
of those little dead eyes staring at you! <shudder>]

You also have to take into consideration line-of-sight obstructions,
esp for "low" (eye-level) mounted sensors. (Up closer to the ceiling
lets you peer OVER things -- but, you tend not to encounter many things
"naturally occurring" at that height.

Think about where the cameras are mounted in banks, shops, etc.

On Sun, 1 Aug 2021 14:18:09 -0700, Don Y <blockedofcourse@foo.invalid>
wrote:

>
>You also have to take into consideration line-of-sight obstructions,
>esp for "low" (eye-level) mounted sensors. (Up closer to the ceiling
>lets you peer OVER things -- but, you tend not to encounter many things
>"naturally occurring" at that height.

Maybe some fancy cornice mouldings with (bat?) gargoyles in the inside
corners? ;-)

--
Best regards,
Spehro Pefhany

On 8/1/2021 2:30 PM, Spehro Pefhany wrote:
> On Sun, 1 Aug 2021 14:18:09 -0700, Don Y <blockedofcourse@foo.invalid>
> wrote:
>
>>
>> You also have to take into consideration line-of-sight obstructions,
>> esp for "low" (eye-level) mounted sensors. (Up closer to the ceiling
>> lets you peer OVER things -- but, you tend not to encounter many things
>> "naturally occurring" at that height.
>
> Maybe some fancy cornice mouldings with (bat?) gargoyles in the inside
> corners? ;-)

Ha! Yeah, I'm sure she'd like that! <rolls eyes>

OTOH, I had planned on "some sort" of decorative treatment for the upper
wall portions. Initially, I'd thought of "rounding" the corners (like
an exaggerated sort of cove base -- does it merit a different name when
mounted at the ceiling??) If I used something purchased (instead of
just installing "curved" wallboard), then I might be able to fit something
to it. And, "service" -- at the worst case -- could just involve removing
the entire length (<frown>).

I keep thinking there has to be some oddball approach I've not yet
considered (like the Ir-through-thin-layer-of-mud suggestion).

In a business/commercial setting, you'd typ have more opportunities
as all sorts of cruft is plastered around a room...

On Sunday, August 1, 2021 at 2:52:55 PM UTC-7, Don Y wrote:

[about mounting/concealing monitor cameras]

> I keep thinking there has to be some oddball approach I've not yet
> considered (like the Ir-through-thin-layer-of-mud suggestion).

How about the use of a sentry critter, with attached camera?
Meerkats, or geese, or prairie dogs will naturally and actively
track anything interesting (to them...).

<https://i2.wp.com/metro.co.uk/wp-content/uploads/2014/03/wpid-meerkat_450x282.jpg?quality=90&strip=all&zoom=1&resize=450%2C282&ssl=1>

On 8/1/2021 6:50 PM, whit3rd wrote:
> On Sunday, August 1, 2021 at 2:52:55 PM UTC-7, Don Y wrote:
>
> [about mounting/concealing monitor cameras]
>
>> I keep thinking there has to be some oddball approach I've not yet
>> considered (like the Ir-through-thin-layer-of-mud suggestion).
>
> How about the use of a sentry critter, with attached camera?
> Meerkats, or geese, or prairie dogs will naturally and actively
> track anything interesting (to them...).

Then I'd need to be able to track the critter! Else, how would
I know the location of the object it was seeing?

Vision, from fixed (known) locations, lets me determine if someone
is in a given area/room, *where* in the room and, possibly, which
direction they are facing.

Presently, I have a general idea as to which room they are in.
And, a general idea as to which *part* of a room (assuming the
room is large enough). I'd like finer-grained resolution
and more "static" unambiguity (e.g., I can't tell which side
of a doorway you are on; and, in some cases, which side of a wall!).

On a sunny day (Sun, 1 Aug 2021 13:42:15 -0700) it happened Don Y
<blockedofcourse@foo.invalid> wrote in <se70v9$iq6$1@dont-email.me>:

>"Hiding" is the wrong word. How about "making cameras less
>of an eyesore"? (I despise having *any* kit visible!)
>
>I'm in the process of updating my "user tracking" system
>to rely on vision (prototype was RF based but that leaves
>open a potential attack vector). So, I need to install
>a boatload of cameras. Most of them *inside* (home,
>commercial establishments, etc.)

All depnds..
I am using several Sony 'super had' analog PAL / NTSC cameras for security
https://www.vision-systems.com/cameras-accessories/article/16745760/sony-introduces-super-had-ii-ccd-sensors-for-security-cameras
only 30$ or so on Aliexpres,
Those things see in the dark without IR just fine (starlight is enough).
But is not HD OK.
Use those in drone for target detection and tracking seeking out vaxers ^H^H^H^H^H^H sorry
small:
https://nl.aliexpress.com/item/32834361928.html
anyways I have the usual movable Chinese with IR LEDs too.

The superhads are small enough to be build in invisible, had one in the front door.
Big 4 channel Chinese analog input security recorder connected via LAN to a raspberry pi 4 and a 3 TB harddisk recording 24/7.

On 01/08/2021 21:42, Don Y wrote:
> "Hiding" is the wrong word.  How about "making cameras less
> of an eyesore"?  (I despise having *any* kit visible!)

You have to decide what features really matter.

Things that are static and the same colour as their background almost
disappear for most people. You can get some pretty small pinhole cameras
but if it is for surveillance then that advantage will be lost if you
also need great panels of IR leds to illuminate the scene. Ideally you
want fast optics and a small compact physical size for this job.

> I just can't see a portion of the solution space that would
> address all of these issues (neglecting cost).  Maybe
> work strictly in Ir and use some sort of semiopaque lens
> to disguise its presence?  (I wonder how transparent a thin
> layer of drywall compound would be to Ir?)

Not very. Cost no object you could perhaps have a thermal band IR imager
working through a thin veneer of pine wood.

--
Regards,
Martin Brown

On 01/08/2021 21:42, Don Y wrote:
> "Hiding" is the wrong word.  How about "making cameras less
> of an eyesore"?  (I despise having *any* kit visible!)

<snip>
>
> I've toyed with some small "spy" cameras with tiny apertures.
> But, they tend to have a fair bit of bulk *behind* the lens.

Laptop camera modules are small and flat. You can readily get 5 MP ones
for a few quid, though 1 MP are more common. They're only a few mm deep
obviously, because they fit in the lid, the PCB is maybe 8mm X 60mm.
They mostly seem to be 3.3V USB which may not suit.

'Laptop camera module' on eBay shows loads.

--
Cheers
Clive

On 8/2/2021 2:09 AM, Martin Brown wrote:
> On 01/08/2021 21:42, Don Y wrote:
>> "Hiding" is the wrong word. How about "making cameras less
>> of an eyesore"? (I despise having *any* kit visible!)
>
> You have to decide what features really matter.

Well, my original/current design didn't depend on visible light
so I didn't even think about that as an issue. :< I can
rationalize NOT needing things to work "in the dark" (as folks
tend to need some amount of light to "do stuff" in a home;
different calculus for a business environment). OTOH, I
can also see NEEDING light to be a drawback for the user.
E.g., visiting the toilet "after dark" shouldn't require the
use of lighting just so the technology can track your location.

But, I suspect I can probably loosen the requirements to just
*tracking* and not user *identification* in those situations.
(maybe not?)

I.e., perhaps settle for a low lux camera and "hope" instead of
trying to incorporate Ir illuminators.

[This is all proof-of-concept so I can always let someone else
"fix" the shortcomings!]

> Things that are static and the same colour as their background almost disappear
> for most people.

Yes, but if you site things high (so they can see over furnishings),
then you tend to be dealing with a uniform wall color/texture
(does anyone use patterned wall paper anymore?) If you site
things lower, you tend to be *in* things that have no fixed
location -- e.g., furnishings.

I have some tiny cameras that are just pinholes (i.e., you can
put them behind a *button* in a sports jacket) so *if* the only
thing visible was the lens aperture, it would just look like
a small dot on the wall.

But, serviceability likely means it's going to be more than a dot
(unless each chosen location can be accessed from behind)

I've seen some homes with PIr sensors mounted in each room.
It looks really tacky. (and, of course, won't solve *my*
problem).

> You can get some pretty small pinhole cameras but if it is for
> surveillance then that advantage will be lost if you also need great panels of
> IR leds to illuminate the scene. Ideally you want fast optics and a small
> compact physical size for this job.

See darkness comments above.

One thing going for me (and against!) is you're typically looking at
short ranges, esp if you shoot a room from all sides. OTOH, short
ranges means field of view becomes a bigger issue.

>> I just can't see a portion of the solution space that would
>> address all of these issues (neglecting cost). Maybe
>> work strictly in Ir and use some sort of semiopaque lens
>> to disguise its presence? (I wonder how transparent a thin
>> layer of drywall compound would be to Ir?)
>
> Not very. Cost no object you could perhaps have a thermal band IR imager
> working through a thin veneer of pine wood.

<frown> I will have to stew on this. The cameras have been a great (almost
universal!) solution in other places -- places where their appearance isn't
typically an aesthetic issue. It would be nice to leverage that approach
in other places.

But, different design criteria tend to require different solutions to
hit the sweet spots...

On Sunday, August 1, 2021 at 3:42:23 PM UTC-5, Don Y wrote:
> "Hiding" is the wrong word. How about "making cameras less
> of an eyesore"? (I despise having *any* kit visible!)
>
> I'm in the process of updating my "user tracking" system
> to rely on vision (prototype was RF based but that leaves
> open a potential attack vector). So, I need to install
> a boatload of cameras. Most of them *inside* (home,
> commercial establishments, etc.)
>
> Just mounting them on walls is easy -- but, makes it look
> like you're in a bank lobby!
>
> I'd thought of down-facing cameras but then I'm limited
> to tracking motion; "features" would tend to be less visible
> (unless perhaps using a 360 degree prism? Likely a
> computational nightmare!). Ideally, I'd like to be able to
> recognize individual users instead of just "presence detect".
>
> The other issue is that they have to be accessible for
> service. The easiest way would be to allow service "from
> the rear" -- another advantage for down-facing (if "the
> rear" is *attic* space)
>
> I've toyed with some small "spy" cameras with tiny apertures.
> But, they tend to have a fair bit of bulk *behind* the lens.
>
> I'll skip over illuminator requirements; worry about if that
> will be an issue *later*!
>
> I just can't see a portion of the solution space that would
> address all of these issues (neglecting cost). Maybe
> work strictly in Ir and use some sort of semiopaque lens
> to disguise its presence? (I wonder how transparent a thin
> layer of drywall compound would be to Ir?)

Maybe something like vintage telephones, model tractors, cars, train locomotives
etc to hide things in plain sight?

On 8/2/2021 4:07 AM, Dean Hoffman wrote:
> On Sunday, August 1, 2021 at 3:42:23 PM UTC-5, Don Y wrote:
>> "Hiding" is the wrong word. How about "making cameras less
>> of an eyesore"? (I despise having *any* kit visible!)
>>
>> I'm in the process of updating my "user tracking" system
>> to rely on vision (prototype was RF based but that leaves
>> open a potential attack vector). So, I need to install
>> a boatload of cameras. Most of them *inside* (home,
>> commercial establishments, etc.)
>>
>> I'd thought of down-facing cameras but then I'm limited
>> to tracking motion; "features" would tend to be less visible
>> (unless perhaps using a 360 degree prism? Likely a
>> computational nightmare!). Ideally, I'd like to be able to
>> recognize individual users instead of just "presence detect".
>>
>> I just can't see a portion of the solution space that would
>> address all of these issues (neglecting cost). Maybe
>> work strictly in Ir and use some sort of semiopaque lens
>> to disguise its presence? (I wonder how transparent a thin
>> layer of drywall compound would be to Ir?)
>
> Maybe something like vintage telephones, model tractors, cars, train locomotives
> etc to hide things in plain sight?

Again, those things can't then be moved. The whole point is to be
able to determine *where* a person is, in a room; who he is; and,
ideally, which direction he is facing.

If someone moves the "vintage telephone" -- or, rotates it so it faces
in a different direction -- then you don't have that positional reference.
E.g., imagine someone moves the "vintage telephone" into a different
room, entirely!

[Yes, it's conceivable that you could do scene analysis to determine
where the camera is CURRENTLY sited. But, that just complicates the
problem. Easier if you *know* it is in a particular location, facing
a particular direction.]

On 8/2/2021 3:11 AM, Clive Arthur wrote:
> On 01/08/2021 21:42, Don Y wrote:
>> "Hiding" is the wrong word. How about "making cameras less
>> of an eyesore"? (I despise having *any* kit visible!)
>
> <snip>
>>
>> I've toyed with some small "spy" cameras with tiny apertures.
>> But, they tend to have a fair bit of bulk *behind* the lens.
>
> Laptop camera modules are small and flat. You can readily get 5 MP ones for a
> few quid, though 1 MP are more common. They're only a few mm deep obviously,
> because they fit in the lid, the PCB is maybe 8mm X 60mm. They mostly seem to
> be 3.3V USB which may not suit.
>
> 'Laptop camera module' on eBay shows loads.

I had a look. The "camera" may be small but there's a fair bit of "support"
that goes along with it; i.e., the PCB isn't very small.

I have some similar to:

<https://www.ebay.com/itm/164953427473>

Again, the camera is small but there's still some bulk "behind it".

And, I'd still have to get the "signal" to something that could analyze
the scene in real-time. WiFi wouldn't cut it as it would be as
vulnerable as my original RF-based scheme.

In my other camera-based applications, I colocate processing resources
with each camera so I can just ship "location data" off instead of
having to run, e.g, RS170 video to "elsewhere" (how many video feeds could
that "elsewhere" handle concurrently?)

I don't think "smaller" is going to be the solution. I think I just
need to find a way to hide most of the "bulk". If there were fewer of them,
I could possibly place them in locations and *deliberately* expose them.
E.g., I have several sirens in the house that are very prominently displayed
(though not easily recognizable as sirens). I've packaged them in a way
that plays off of other "appliances" that *can't* be hidden (e.g.,
smoke/CO detectors). But, I have a lot more leeway in where to site
sirens as they work equally well if sited *here* or *there*...

On 8/1/2021 1:42 PM, Don Y wrote:
> "Hiding" is the wrong word. How about "making cameras less
> of an eyesore"? (I despise having *any* kit visible!)
>
> I'm in the process of updating my "user tracking" system
> to rely on vision (prototype was RF based but that leaves
> open a potential attack vector). So, I need to install
> a boatload of cameras. Most of them *inside* (home,
> commercial establishments, etc.)
>
> Just mounting them on walls is easy -- but, makes it look
> like you're in a bank lobby!
>
> The other issue is that they have to be accessible for
> service. The easiest way would be to allow service "from
> the rear" -- another advantage for down-facing (if "the
> rear" is *attic* space)
>
> I'll skip over illuminator requirements; worry about if that
> will be an issue *later*!

From discussions with friends/colleagues, it looks like I'll
have to roll-my-own (package). But, even prototype quantities
will be high enough to justify the NRE...

Spehro Pefhany wrote:
> On Sun, 1 Aug 2021 13:42:15 -0700, Don Y <blockedofcourse@foo.invalid>
> wrote:
>
>> "Hiding" is the wrong word. How about "making cameras less
>> of an eyesore"? (I despise having *any* kit visible!)
>>
>> I'm in the process of updating my "user tracking" system
>> to rely on vision (prototype was RF based but that leaves
>> open a potential attack vector). So, I need to install
>> a boatload of cameras. Most of them *inside* (home,
>> commercial establishments, etc.)
>>
>> Just mounting them on walls is easy -- but, makes it look
>> like you're in a bank lobby!
>>
>> I'd thought of down-facing cameras but then I'm limited
>> to tracking motion; "features" would tend to be less visible
>> (unless perhaps using a 360 degree prism? Likely a
>> computational nightmare!). Ideally, I'd like to be able to
>> recognize individual users instead of just "presence detect".
>>
>> The other issue is that they have to be accessible for
>> service. The easiest way would be to allow service "from
>> the rear" -- another advantage for down-facing (if "the
>> rear" is *attic* space)
>>
>> I've toyed with some small "spy" cameras with tiny apertures.
>> But, they tend to have a fair bit of bulk *behind* the lens.
>>
>> I'll skip over illuminator requirements; worry about if that
>> will be an issue *later*!
>>
>> I just can't see a portion of the solution space that would
>> address all of these issues (neglecting cost). Maybe
>> work strictly in Ir and use some sort of semiopaque lens
>> to disguise its presence? (I wonder how transparent a thin
>> layer of drywall compound would be to Ir?)
>
> Have a look at some of the spy camera setups if you're interested in
> that sort of thing- mounting in smoke detectors, in wall power
> outlets, in coat hooks, in clock radios where there is natural space
> behind the aperture. I'm sure paintings are another possibility.
>
> Eg. https://amzn.to/3C6Hi15
>
Fortunately all cameras work as cats' eye reflectors, so they're not
hard to spot with the right laser scanning kit.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

On 8/2/2021 6:10 PM, Phil Hobbs wrote:
> Spehro Pefhany wrote:
>> On Sun, 1 Aug 2021 13:42:15 -0700, Don Y <blockedofcourse@foo.invalid>
>> wrote:
>>
>>> "Hiding" is the wrong word. How about "making cameras less
>>> of an eyesore"? (I despise having *any* kit visible!)
>>>
>>> I'm in the process of updating my "user tracking" system
>>> to rely on vision (prototype was RF based but that leaves
>>> open a potential attack vector). So, I need to install
>>> a boatload of cameras. Most of them *inside* (home,
>>> commercial establishments, etc.)
>>>
>>> Just mounting them on walls is easy -- but, makes it look
>>> like you're in a bank lobby!
>>>
>>> I'd thought of down-facing cameras but then I'm limited
>>> to tracking motion; "features" would tend to be less visible
>>> (unless perhaps using a 360 degree prism? Likely a
>>> computational nightmare!). Ideally, I'd like to be able to
>>> recognize individual users instead of just "presence detect".
>>>
>>> The other issue is that they have to be accessible for
>>> service. The easiest way would be to allow service "from
>>> the rear" -- another advantage for down-facing (if "the
>>> rear" is *attic* space)
>>>
>>> I've toyed with some small "spy" cameras with tiny apertures.
>>> But, they tend to have a fair bit of bulk *behind* the lens.
>>>
>>> I'll skip over illuminator requirements; worry about if that
>>> will be an issue *later*!
>>>
>>> I just can't see a portion of the solution space that would
>>> address all of these issues (neglecting cost). Maybe
>>> work strictly in Ir and use some sort of semiopaque lens
>>> to disguise its presence? (I wonder how transparent a thin
>>> layer of drywall compound would be to Ir?)
>>
>> Have a look at some of the spy camera setups if you're interested in
>> that sort of thing- mounting in smoke detectors, in wall power
>> outlets, in coat hooks, in clock radios where there is natural space
>> behind the aperture. I'm sure paintings are another possibility.
>>
>> Eg. https://amzn.to/3C6Hi15
>>
> Fortunately all cameras work as cats' eye reflectors, so they're not hard to
> spot with the right laser scanning kit.

As stated, the goal isn't to "hide" them but, rather, to make them less of
an eyesore.

Spend a few moments and you can clearly see and find ALL of them.
But, they don't want to be the thing you remember about a "space".
I'm not interested in catching you (a resident/employee!) doing something
"illegal"; rather, I just want to know where you are.

By contrast, walk into a bank and the cameras are very deliberately
visible. They aren't concerned with aesthetics (it's a BUSINESS)
and want to keep pressure on you to know that "someone is watching".

When you go to the self-check at most stores, there's a video
DISPLAY showing you your image as you are conducting your
business. This, despite a sign informing you of surveillance
cameras in use. Clearly, there are other cameras scattered
around the store -- almost all of them are pretty visible if
you just LOOK for them. Highlighting them is intentional;
to keep you honest.

[This is why some stores have greeters -- because a "human interaction"
with the store decreases the likelihood of theft (yeah, there are
people who study this sort of stuff)]

On Monday, August 2, 2021 at 6:28:03 PM UTC-7, Don Y wrote:

> >> On Sun, 1 Aug 2021 13:42:15 -0700, Don Y <blocked...@foo.invalid>
> >> wrote:

> >>> Just mounting them on walls is easy -- but, makes it look
> >>> like you're in a bank lobby!

> As stated, the goal isn't to "hide" them but, rather, to make them less of
> an eyesore.

How about using telephoto lenses, and a convex mirror on the ceiling? The cameras can be
above eye level, and concealed from direct sight, but still see the convex mirror
(which spreads the view out to cover much of the room).

For extra credit, correct for visual field curvature (project a grid onto the floor and
use that coordinate system to train the correction, perhaps). A second mirror, or
software, can correct for image inversion.

Few humans will bother to analyze the curved images in the silvery blob above, so
the camera(s) can remain unnoticed. They'd be apparently distant objects in any case.

On 8/2/2021 7:01 PM, whit3rd wrote:
> On Monday, August 2, 2021 at 6:28:03 PM UTC-7, Don Y wrote:
>
>>>> On Sun, 1 Aug 2021 13:42:15 -0700, Don Y <blocked...@foo.invalid>
>>>> wrote:
>
>>>>> Just mounting them on walls is easy -- but, makes it look
>>>>> like you're in a bank lobby!
>
>> As stated, the goal isn't to "hide" them but, rather, to make them less of
>> an eyesore.
>
> How about using telephoto lenses, and a convex mirror on the ceiling? The cameras can be
> above eye level, and concealed from direct sight, but still see the convex mirror
> (which spreads the view out to cover much of the room).

How do you "conceal(ed) from direct sight"? Perhaps if you had indirect
lighting and could hide the cameras up in the sconce?

> For extra credit, correct for visual field curvature (project a grid onto the floor and
> use that coordinate system to train the correction, perhaps). A second mirror, or
> software, can correct for image inversion.
>
> Few humans will bother to analyze the curved images in the silvery blob above, so
> the camera(s) can remain unnoticed. They'd be apparently distant objects in any case.

I think that would be even *more* "noticeable" than a bunch of ~5mm "shiny
spots" on the walls! And, I learned there's a fair bit of processing involved
in trying to compensate for optics (I have a HAL9000 at my front door that
surveils and recognizes visitors. My first implementation used some junk
optics -- intended more to represent the physical characteristics of the
hollywood piece: "I can always compensate in software..." Bad move!)

I think I have a good packaging solution that will let me leverage the existing
electronics. I'll have to see if I can, also, apply it to the cameras in the
garage (doubtful that I an apply it to the cameras on the outside of the house
as they are optimized for different operating and detection conditions)

On Monday, August 2, 2021 at 7:37:06 PM UTC-7, Don Y wrote:
> On 8/2/2021 7:01 PM, whit3rd wrote:
> > On Monday, August 2, 2021 at 6:28:03 PM UTC-7, Don Y wrote:
> >
> >>>> On Sun, 1 Aug 2021 13:42:15 -0700, Don Y <blocked...@foo.invalid>
> >>>> wrote:
> >
> >>>>> Just mounting them on walls is easy -- but, makes it look
> >>>>> like you're in a bank lobby!
> >
> >> As stated, the goal isn't to "hide" them but, rather, to make them less of
> >> an eyesore.
> >
> > How about using telephoto lenses, and a convex mirror on the ceiling? The cameras can be
> > above eye level, and concealed from direct sight, but still see the convex mirror
> > (which spreads the view out to cover much of the room).

> How do you "conceal(ed) from direct sight"? Perhaps if you had indirect
> lighting and could hide the cameras up in the sconce?

The camera only has to look UP, so put a decorative bowl on a high shelf, and
put the camera in the bowl. Only the reflection in the mirror is in line-of-sight.

On 8/2/2021 8:31 PM, whit3rd wrote:
> On Monday, August 2, 2021 at 7:37:06 PM UTC-7, Don Y wrote:
>> On 8/2/2021 7:01 PM, whit3rd wrote:
>>> On Monday, August 2, 2021 at 6:28:03 PM UTC-7, Don Y wrote:
>>>
>>>>>> On Sun, 1 Aug 2021 13:42:15 -0700, Don Y <blocked...@foo.invalid>
>>>>>> wrote:
>>>
>>>>>>> Just mounting them on walls is easy -- but, makes it look
>>>>>>> like you're in a bank lobby!
>>>
>>>> As stated, the goal isn't to "hide" them but, rather, to make them less of
>>>> an eyesore.
>>>
>>> How about using telephoto lenses, and a convex mirror on the ceiling? The cameras can be
>>> above eye level, and concealed from direct sight, but still see the convex mirror
>>> (which spreads the view out to cover much of the room).
>
>> How do you "conceal(ed) from direct sight"? Perhaps if you had indirect
>> lighting and could hide the cameras up in the sconce?
>
> The camera only has to look UP, so put a decorative bowl on a high shelf, and
> put the camera in the bowl. Only the reflection in the mirror is in line-of-sight.

Shelf? What's that? :>

We have a shelf in each closet. And, I have a wall of bookshelves in
my office. And, one of the pantries has shelves, floor to ceiling.

But, no "shelves" anywhere else in the house! Walls are adorned with
artwork. No "nick-nacks" to dust -- the sorts of things you'd put
*on* shelves, etc.