I'm getting into an active mode on the new alarm system. See also
New Alarm System
, which has earlier
design plans.
This month I got the ZoneMinder camera recording software working, using a little USB webcam for testing, and I purchased and tested a Wyze Cam Pan ($38). But it's obvious that I need to go back to talking about and deciding several key aspects of the alarm system, which would normally come before the action items.
A specific event put a bee in our bonnet, alarm-wise. We were in another city on vacation, and the Napco Gemini II intrusion alarm detected a threat. At that time it was professionally monitored. They contacted us by phone, and we decided that calling the police would be a good idea, which they did. There was no actual intrusion, and we got stuck with a $350 bill for a false alarm.
The bill is annoying but in the grand scheme of alarm selection it's a small issue. But its real significance is in making clear an inadequacy of our existing security system. Here's what we think should actually have happened:
The intrusion alarm should not have gone off for a false positive, which we're attributing as just another symptom of the bit rot, but false alarms are inevitable in any security system. In fact we can expect that the majority of events will be real but ignorable events, e.g. I walk through the living room or out the front door. The majority of remainder events will be explainable false positives, e.g. a cat scratches on our glass door, setting off the glass break sensor, or a large palm frond drops in the wind, setting off a security camera's motion detect algorithm. A very small minority of events will be true positives, e.g. a burglar trying or succeeding to enter.
Whichever subsystem detected the threat, it should trigger a push notification so we know promptly to pay attention. How to make that happen is complicated.
We would want to view the target area and make a judgment whether we saw any nefarious activity, e.g. a broken window. We don't have that capability now, and by not having it, we blew a false positive into a big traumatic scene. For this kind of service it would really help to have pan-tilt-zoom (PTZ) control of the camera. Two-way audio to harass the burglar sounds sexy but we are unlikely to ever catch a burglar on screen in real time, just on delayed video clips.
So what capabilities do we want?
One of the most important capabilities is integration: we would like a coordinated response to issues on any subsystem, and particularly, we want to do user training on just one interface, not a gaggle of very loosely federated subsystems.
Push notification to a mobile app, for use when we are travelling.
Window and door open detection. (Have, wired.) We rely on this to verify that the house is ready for going to sleep.
Motion detectors in various rooms (non-imaging). (Have.)
A small number of cameras with PTZ. (New.) 2-way audio is not a high priority.
We have two glass break sensors (audio).
Smoke and CO sensor integration. We have the sensors but they are audio only.
Water leak sensor. (New.) When the house was being built there was a water leak, so we're sensitive to the issue.
We have an audio doorbell intercom, which is very useful. We could consider a modern audio-video doorbell.
A Bluetooth door lock is very sexy, but we would probably end up rejecting it due to worries about multiple failure modes.
Alternatives are:
Full cloud integration. The Honeywell thermostats are an example. Pros and cons:
Using their base station with modest cloud involvement. The Obihai ATA for VoIP telephony is sort of an example. Pros and cons:
Locally integrated solution. Presently ZoneMinder is the front runner for security camera software. Pros and cons:
A gaggle of systems. The intrusion alarm only does intrusion, the cameras independently do only video, the smoke alarms only do smoke, etc. Of course we're totally in that mode already. Pros and cons:
What features do I actually want in my camera(s)? For my use case I'm looking at detecting a breakin by other means, and using the camera to determine if it is, or isn't, a false alarm.
Streaming video is preferred over still images.
Visible and infrared light (short wave) each have their own advantages. Visible (in color) is most useful for recognizing false alarms with no burglar, but at night it would need a lamp, which in case of a false alarm would wake us up. Infrared has complementary strong and weak points; an infrared illuminator is commonly included on such cameras. Some cameras can do both.
We're going to have to play it by ear on the field of view. Extreme wide angle seems a bit extreme. 90° feels about right (I'm guessing).
Sensor resolution is a guessing game. Likely 720p is sufficient. Steroidal resolution is not appreciated; we have to allow up to 3 cameras sending at once, and we have to store the images. Onboard video compression would be a big plus.
It's not acceptable to stream video continuously. One feasible mode is onboard motion detection, although my plans lean toward detecting the threat by other means and then recording from the cameras. It's important to have a ring buffer in which video is saved from before the alarm occurs, so the moment of entry can be captured, probably with the best view of the burglar (if not a false alarm).
Pan-tilt-zoom capability likely would be helpful.
While I would prefer to communicate by Ethernet, almost certainly I'll be stuck with Wi-Fi because I can't run wires to the camera.
For the same reason, power will be from batteries. It's essential to be able to read out the battery charge level.
Some cameras have a microphone and a speaker, for harassing the burglar from a remote site. Other people use this feature for monitoring a baby or children or pets (that respond to e.g. "stop exercising your claws on our couch!")
It's interesting to read the terms of service for the Wyze Cam. They explicitly state that it is not to be used for security or life safety purposes (don't sue us if it works sub-optimally when you are burglarized), and they explicitly mention that a sophisticated attacker can jam wireless signals. All the product hype shows monitoring babies or pets.
The first step is going to be to obtain one security camera and to set it up indoors with the eventual monitoring software, expected to be ZoneMinder.
List of cameras included as parts of starter kits mentioned in the previous document:
Abode
combination infrared motion detector and camera.
The camera, when triggered, takes 640x480px color photos; it has
flash (visible light). 90° field of view. Power from coin cells.
Communicates by Ethernet or Wi-Fi.
A different product is a streaming video camera (720p).
LifeShield
offers a streaming video camera at 720p 30FPS.
66° field of view. It has an infrared illuminator, so the image must
be shortwave infrared. Communicates by Ethernet or Wi-Fi. It includes
motion and sound detection. It has a SD card slot for autonomous video
storage. The stand can pan and tilt manually.
Other vendors offer cameras but not in their starter kits, so the cameras were not reviewed in these articles.
Google search for home security camera
— there are a lot
advertised. I'm starting with this
review from Safewise, Best Home Security Cameras for 2019
by Rebecca Edwards, 2019-11-27.
Arlo Pro (by Netgear): These reviewers (and others) consider this to be a top of the line camera. But it has two disadvantages: you have to use their base station, which is a showstopper, and it costs $249.
Wyze Cam Pan:
Ring Spotlight Camera Wireless:
The authors list quite a number of other cameras but did not do detailed reviews on them.
Another interesting article:
Wyze Cam: The cheapest security camera I've ever tested might be my favorite
by Megan Wollerton, 2019-08-29.
This is the bare bones member of Wyze's product line, having a manually
tiltable base. Other than that, the specs seem similar to the Wyze Cam Pan
reviewed above: 14 days cloud storage, free alerts, and a micro-SD slot if
you want autonomous recording. She thinks it's a very good camera,
particularly at this price.
ZoneMinder has full integration, so they say, with the Wyze Cam series. This is the one I'm going to try out, as the first candidate.
This feature list is from the Wyse web site and jimc has kind of commingled the Pan and non-Pan cameras. Some specs are slightly different between them and I didn't list them separately if minor.
skinfor the Wyze Cam Pan.
Which cameras are supported by ZoneMinder? Basically, a lot of them. Mostly I'm cueing on cameras mentioned above. The start point for this is the ZoneMinder Hardware Compatibility List on their wiki.
Most modern IP cameras are ONVIF compliant, meaning ZoneMinder can deal with them.
fullycompatible.
I was able to make it do these features, controlling it from the mobile app:
I installed the Wyze mobile app and created an account.
I followed the setup instructions, joined the wireless net, updated the firmware which as usual was way back version.
Live stream from the camera, viewed with the mobile app. There's about 0.5sec latency between something happening in the scene, and seeing it on the mobile app.
Pan and tilt (yaw and pitch). In the mobile app you hit a virtual corner of the control ring and it steps a preset distance in that direction.
Infrared sensor and illuminator: viewing with no visible light gave a useful image. 30ft range may be kind of an overestimate.
Audio: the speaker icon labelled Sound, if clicked, lets you listen to what the camera hears. The microphone icon labelled Voice lets you speak to the burglar or pet. It's not very loud.
Power saving on inactivity: it draws about 3W when you're actively doing tests, but under 1W (reads as 0) when the scene is static.
Motion detection: Moving objects are highlighted with a targeting box, as seen on the mobile app. I'm assuming that the camera does this onboard, without proof.