New home surveillance set-up

[codybby]

Hi everyone! I'm new to the forum and new to home surveillance. I'm about to the start construction on a new home and having a way to keep an eye on things are high on the list of things I want, unfortunately that task is left up to me to figure out and needless to say I'm a noob.
I've looked at some camera packages and could probably buy one of those and be perfectly fine but If the tech world has taught me anything it's that you can do more for less $ when building yourself and get better performance.

I'd like to have 4 to 8+ cameras setup to my PC for viewing but also have ability to be viewed from my phone and such when I'm away. A good 4K camera would be nice if I can power it via PoE, I don't want to run anything more than ethernet cables to the cameras.

 

[Zorac]

you don't really have to worry about to much detail right now, just make sure you run enough cable everywhere. and when you think you have enough runs, add some more! always seem to need another run.... when you take possession of the house in 6 months, there might be newer models, or at the very least the existing ones should be a little cheaper.

 

[codybby]

you don't really have to worry about to much detail right now, just make sure you run enough cable everywhere. and when you think you have enough runs, add some more! always seem to need another run.... when you take possession of the house in 6 months, there might be newer models, or at the very least the existing ones should be a little cheaper.

That's an excellent point. I was just going to run conduits to the camera locations so I can change cables as needed though at first I'm just gonna put cat6 in.

 

[Fastb]

while building/constructon, consider laying pipe (pvc or cheap irrigation tube, ie: funny pipe) to places away from the house. cams outside aimed at people approaching the house give a different perspective. Maybe more impotantly, they let you connect dependable sensors to your system.
cams have MD, tripwire ans such, but the false alarms turn into a nuisance. a true sensor (pir which trips with body heat, not moving shadows and shrubery) is far superior for security. Security (to alert you to trespassers) , and cams (to tell you what happned, usually after-the-fact due to undependable alarms) are two masters. one solution can't usually serve both masters

 

[fenderman]

You dont want a 4k camera....they will be terrible at night and low light unless you will spend serious cash..

 

[Zorac]

That's an excellent point. I was just going to run conduits to the camera locations so I can change cables as needed though at first I'm just gonna put cat6 in.

conduit is even better! be pricey though and your builder might give you grief, they will want to just drill the holes in the studs and pull the wire (not my area, but if your supplying the wire, get stuff that is rated for risers and no cca). builders don't normally do conduit and the electricians will probably overcharge to do it as a change, have the talk with them though and see what pricing they give you on the options. give some though about where you want to run your cable too and perhaps allow for a small wall mounted netshelter or shelf (for router/switch/modem/poe injection/voip/nas/etc). phone/cable should be brought to the same place for your modem.

my house is old, so i pulled all my wire (cat6 as well) through my drop ceiling in the basement and attic which worked ok. no one was thinking network cable in 1980 unfortunately!

 

[j4co]

I am in the same situation of building a new house. So i use a layout map of the house and check where the entry points are, where the cars are parked etc. You can than measure the distances and than you would need to calculate which field of view and what pixel resolution you need at the position you need to identify persons.
after that you buy camera's with the specifications you need.

For example my home:


Text is all in Dutch as i need it for the builder later

the front is bottum and that is simple. Left side will be a gate, so needs a cam watching that. Right side is car parking garage door and house entry, so need cam there. Front of house 2 more so i can watch front of house, or closer to cars from other angle.
back of the house worries me, 2 doors so 2 cameras, but large garden, so high mounted ptz with some detectors covering the garden.

just an example..

 

[nats]

cams have MD, tripwire ans such, but the false alarms turn into a nuisance. a true sensor (pir which trips with body heat, not moving shadows and shrubery) is far superior for security. Security (to alert you to trespassers) , and cams (to tell you what happned, usually after-the-fact due to undependable alarms) are two masters. one solution can't usually serve both masters

This is not true anymore, there are services which will connect to your cameras and issue relevant alerts, like only alerting when a human is detected.

The way I see it, if you have a camera looking at something happening, you should be alerted in realtime.

 

[Fastb]

nats,

the "services" you mention;
1) is that a "monitoring" service by a 3rd party? My comment was related to self monitored systems.
2)) only alert when "human" detected; Sounds like the traditional paid monitoring service approach. A PIR sensor will trip on IR motion, ie a warm body. Human, dog, deer, etc. To reach a conclusion the intruder is "human" has required a thinking human to make that conclusion...

Lastly, my comments pertain to exterior applications, a tougher application than indoors, where defining MD zones and MD object size can eliminate the dog & cat from tripping a false alarm. those solutions don't translate well to monitoring an open & large back yard.

Can you elaborate on your intriguing statement about "services which will connect to your cameras"

Fastb

 

[nats]

1) is that a "monitoring" service by a 3rd party? My comment was related to self monitored systems.

Yes and no. There is the option of the more traditional 3rd party monitoring, but also something in between. A service based solely in software which radically improves the quality of the alerts a user would get, and allow the customer to react in real time. So essentially it would still be a self managed solution, but where you only get relevant alerts.

The more traditional solution basically uses the same software internally. It is just a way to offer a more affordable option if someone wants it.

2)) only alert when "human" detected; Sounds like the traditional paid monitoring service approach. A PIR sensor will trip on IR motion, ie a warm body. Human, dog, deer, etc. To reach a conclusion the intruder is "human" has required a thinking human to make that conclusion...
Lastly, my comments pertain to exterior applications, a tougher application than indoors, where defining MD zones and MD object size can eliminate the dog & cat from tripping a false alarm. those solutions don't translate well to monitoring an open & large back yard.

Recently, the tech advances have made it posible to do human detection without a human. Fun fact, this year a computer beat humans in that task exactly. There is no reason for that tech to not be applied to helping out on video security today.

Below is a video of an outdoors application for the software only solution.

Can you elaborate on your intriguing statement about "services which will connect to your cameras"

The service is basically a cloud base solution which connects to a video stream of your IP cameras or DVR and monitors them for you. Sends alerts to your phone and SMS or phone calls if you want them. Basically if you have a camera which can expose a video stream externally (port forwarding might be required) then you can connect it.

The same analysis identifies other objects like cars, dogs, birds, etc. But the focus right now is on humans and intruder detection.

 

[jasauders]

Hate to be blunt, but there's something about the idea of a 3rd party service "monitoring" my video feeds from outside my network that freaks the hell out of me.

 

[Fastb]

nats,

interesting video.

I gotta say, the beach pathway application seems like a "medium-tough" application. Specifically, a PIR motion detector, or a dual technology sensor (PIR + radar) may perform similarly. ie: little or no false alerts. That scene didn't have moving foliage, which trips up camera based MD. And any moving body on that beach side path is highly likely to be a human.

My dahua cams will draw boxes around moving objects (to support IVS tripwire and intrusion zone features) Your system ignored dogs, however. The Dahua IVS supports setting parameters for min & max object size, which effectively filters out small to medium dogs. If I filter out a large dog, I could miss a small person. And the Dahua object size definition doesn't scale with perspective - a person at a distance takes as mnay pixels as a dog near the camera. I think it's all pixel-count based, not a calculated object size, based on inferred distance.

I think in a "really-tough" application, your system would really shine, compared to an application that only used Dahua IVS. The "really tough" application might include dancing shadows form blowing tree limbs, moving shrubbery, moths flying by the camera, all while people are walking by.

So for an application that is limited to "camera only", that image processing is nice, and would reduce false alarms. However, the "camera only" constraint rigs the contest a bit. It rules out a [camera + PIR], or (camera + dual tech sensor + laser line). An evaluation would consider equipment cost, monthly costs (including the cost of streaming the live video + human detection image processing fees), installation ease/difficulty, and lastly, performance.

I know laser lines and PIR sensors operate well in pitch black. How well does the "Human Detection" image processing do with b&w With low light? Will it miss Trick-or-Treat vistors dressed like ghosts? (head, arms & legs hidden by a bedsheet)

Nats, your technology seems cool - I don't mean to sound negative. For some installations & applications, it's ideal. We may this image processing built into cameras or NVRs in the years to come.

Fastb

 

[nats]

I think in a "really-tough" application, your system would really shine, compared to an application that only used Dahua IVS. The "really tough" application might include dancing shadows form blowing tree limbs, moving shrubbery, moths flying by the camera, all while people are walking by.

The tech is really robust on trees, changing lights, car lights, and bugs, etc. The detection at night is as good as the IR leds. As with any installation, IR illumination should be adequate for the monitored area. So far we have not detected any noticeable effect on detection for sunlight vs IR illumination.

So for an application that is limited to "camera only", that image processing is nice, and would reduce false alarms. However, the "camera only" constraint rigs the contest a bit. It rules out a [camera + PIR], or (camera + dual tech sensor + laser line). An evaluation would consider equipment cost, monthly costs (including the cost of streaming the live video + human detection image processing fees), installation ease/difficulty, and lastly, performance.

The same service will allow for your alarms to evolve in all kinds of ways though. Like recognizing your face and not alerting. Or setting alarms for very specific things for being alerted.

Also, the camera-only setup, makes for a simplified setup which is easier to install and maintain. Simplifying the requirements and costs for quality surveillance is also something that I think is very interesting.

Nats, your technology seems cool - I don't mean to sound negative. For some installations & applications, it's ideal. We may this image processing built into cameras or NVRs in the years to come.

This processing will definitely make it into the hardware. However, for all the installations already out there, this is an option to get it without changing your whole setup.

Thanks for your feedback.

 

[Janek]

Hi guys,

Please allow me to jump in and ask you for some input.
My business partner and I are / have developed a software that can run on almost any camera system and can provide Visual Fire/Smoke detection and alerts.
We are in the process of licensing this software, but would like to get it out into the open market for testing.
So i wonder if any of you would be interested in testing such software and possibly participate on its further development and even possibly on its deployment in the US.
We are based out of Europe but would like to explore the massive potential of this software in the US.

Please let me know if any of you would be willing to give it a shot and do some testing and simulations on your camera system.
im not sure how to send or receive PMs on this board, but if any of you are interested just let me know and i can give you my email address for further communication.
hopefully im not breaking any board rules here with this post.
if yes, i apologize in advance, but i thought what better way than to reach out to the camera experts on this board

thanks .
JD