Hacker News Reader

Comment by sliken | Hacker News Reader

For those that want to avoid such sillyness, Reolink sells relatively cheap cameras. Rated for out doors, power over ethernet, $50-$60 per camera, and includes a microphone.

They can easily be connected to zone minder, or any software that can take a rtsp:// URL. Even handles motion detection for specific areas of camera, so you can include the driveway but exclude the sidewalk. You can have it email or upload videos... without access to any reolink related cloud.

So you could easily put them in production with zero network access and let something you control notify you with images or video clips for any activity.

There's numerous cheap products, but the reolink seems to be one of the better ones that play well with others and doesn't require any WAN network access.

Ubiquiti and Axis also have some very nice products, but generally are more expensive.

View on HN

Slightly OT, but I'm curious to understand the use-case for indoor/outdoor home-security cameras. Yes, it sounds pretty awesome to "catch" someone, or trigger a motion alarm, but does it work in practice? Like, is it hooked to the police or to a security firm, or what? Because someone with a face mask is unidentifiable anyway, and even if you can identify some feature, a burglar can leave with whatever valuables, or cause damage... So it's not preventing the crime, but might marginally aid in after-the-fact investigation? so is it worth it?

I'm probably not a good example, I live in a small apartment, don't have any valuables besides passports and loose change. I just don't see much of a point for having cameras. Instead I focus on deterring casual burglars by using semi-intelligent timer-based lights plus a simple radio tuned to a station with lots of talking when we're away from home. Perhaps a security cam has a deterring aspect, but not sure of the legality of it in Germany where I live, and I guess a dummy camera would achieve a similar effect as a real one?

What am I missing? :)

Well my motivation started when I found an uncoiled coat hanger on top of my car. So I bought two cameras (both with IR illumination) and pointed them to the two sides of my car. Even a casual observer will notice the two cameras (with 2 rings of glowing red lights) pointed at them when they enter my driveway. I figure it's a $120 worth of prevention... hopefully.

The cameras can read the license numbers of cars going up and down my dead end street, and have a NTP synced time stamp on them. So I have reported license plate numbers to the police a few times when various neighborhood issues have come up.

But no, I don't think it will magically arrest someone breaking into my house. But they are still pretty useful. Did someone leave a package on my front door? Is my kids friend's parent here for pickup/drop off? Did the garbage truck get here yet? WTF is that loud truck noise out front? Is that knock at the door someone with a clip board? Or someone trying to get me to believe in their god?

One thing it does help, is that at least in my area it's common for a thief to knock to see if anyone is home, then break down that door or force a rear door to enter the house is nobody answers. So being able to see and even respond to the knocker might well prevent a breakin, even if you aren't physically there.

Chances are low to identify someone (for the time being, but increased video quality and facial recog. vs. a simple driver's license/ID database may change that soon). Obviously like you said, nothing you can do with a face mask.

But at least if you know something is happening in real-time, motion triggers being the most important, there is at least some chance to intervene or call the police depending on the circumstances. Even if you can't stop them, speed is key, and if they are on foot they could be caught in the area afterwards with vague id such as clothing only. Capturing a license plate is probably optimal though, if they aren't on foot. Cameras are very good these days - HD and higher 30/60fps video - not the pixelated/blocky/useless recordings of 10 years ago.

Also good to know ahead of time if there is anyone snooping around looking for easy targets (as small-time burglars do). Some systems do also have monitoring services.

Otherwise yeah, cameras are at best a deterrent and source of evidence, nothing more as a passive device. Best security system I've ever had was a German Shepherd. Fully autonomous deterrent + active response if necessary :)

The cameras are also a great deterrent just for being there. After installing some on our vacation house, our gardener told us there's almost never junk flyers anymore.

I like the Ubiquiti cameras because unifi handles the webrtc handshake so you can run your cameras local and see them from anywhere while keeping data on-prem.

It's a little annoying Protect doesn't have a software install but the legacy unifi video still work fine. I put in ~5 cameras over the last few weeks and been really happy with them.

FWIW: I just upgraded to UniFi Protect (with the CloudKey Gen2). Although it's annoying I don't get to run the software on my own server anymore, I've found the Protect app for iOS is much more polished and is much more responsive than the older UniFi Video app.

Good to hear. I led the UniFi Protect effort at Ubiquiti. Responsiveness and fast time-to-video were our top UX priorities for the system. The WebRTC direct-connect and bypassing cloud servers were big factors in delivering that responsiveness. The added security and lack of monthly fees was another major bonus.

While I have you here, please consider the option to install the software on-prem.

I was this || close to not going unifi because it wasn't clear what the longtime support for unifi video was. I'm still not super-happy with that aspect as it's not clear if the hardware is tied to protect or not.

AFAIK the cameras still output standard RTSP video, so you could still use those streams in other software.

The problem I have with protect is I already have a full blown Proxmox setup with mirrored ZFS drives that I can easily expand.

I lose that if I switch to their custom hardware. I've also heard it doesn't scale well to multiple users or larger numbers of cameras.

Having thrown countless hours into zoneminder I had to give up. It's just too buggy. Does anyone have any other alternatives they've got to work? Open Source or otherwise

Try shinobi or motioneye if you want open source.

There's quite a few solutions in this space: free, freemium, and commercial.

I don't recommend Shinobi after using it for a few months. It's far from finished, can't do most basic stuff and the interface is horrible.

I saw a commercial software solution a number of months ago that used some kind of machine learning or AI to reduce false positives on motion detection but now I can't find it. Are you familiar with any software that does this?

Look into Milestone XProtect, super solid, have a free version.

I tried zoneminder in earnest and was severely disappointed. I've been running Ubiquity Video in a container on my local server, which is pretty great as a simple NVR, but _only_ works with their cameras, which won't work for me.

Blue Iris seems to be the overall favorite, and it's pretty inexpensive considering the fact that it seems to be so well regarded. I don't have Windows on my network, and I was strongly considering running BI on KVM, but I don't think my old server is beefy enough to reliably run windows with video software. I was also looking into building a box for it, but it turns out buying a used Optiplex on Ebay was cheaper and should be more than enough for tracking a few cameras.

I'll be setting it up this weekend, so I don't have much of my own experience to share, yet.

Synology Surveillance Station is pretty nice, but closed source.

Another option is from Amcrest and is built as a doorbell. I got it because it doesn’t require a cloud connection to function and plugs in easily to Home Assistant. Easy install with existing wiring and uses WiFi for data.

Note: one disappointment was that the app automatically reached out to an Amcrest server by default (I assume) on the assumption that everyone wants access to their home doorbell cam from outside the network ... I could not find a setting in the app so just took care of it the usual way—blocked it with the firewall. Regardless, it still works in this config.

If you are home, can you virtually answer the door with the amcrest without using any cloud services? From a desktop browser? From an android phone?

As far as I can tell, no.

Ever since I blocked the outgoing connection via firewall, the app now says my camera is offline. That being said, the camera feed is still working in Home Assistant, which is all I wanted. If you were using HA and wanted more functionality (audio, etc), it does appear to be available.

https://www.home-assistant.io/integrations/amcrest/

Wyze cams now have an RTSP firmware as well and cost about $25 for stationary, and $40 for PTZ.

I was looking into Wyze right about the time the server leak happened [0] so I moved on. But they do have good prices.

[0] https://www.theverge.com/2019/12/30/21042974/wyze-server-bre...

I've been using Wyze to keep tabs on a new pet while I'm not at home. It is very convenient, but I was pretty worried about privacy, even before that came out.

At $25 to replace, it's probably worth risking bricking it to try out alternative firmware: https://github.com/openipcamera/openipc-firmware

This repo seems unmaintained. This project is more active: https://github.com/EliasKotlyar/Xiaomi-Dafang-Hacks

All I want is an outdoor rated rpi camera case with built-in power supply - motioneyeos will take care of the rest.

Just 3d print a case and use silicone to waterproof all electronics and case.

Maybe even use an esp32 cam board instead of a whole raspberry pi:

https://randomnerdtutorials.com/esp32-cam-video-streaming-we...

https://www.thingiverse.com/thing:3996434

Not sure what you mean by built-in power supply. A battery? Or just a weatherproofed jack to connect a cable?

An in-case mains AC/DC converter.

That's probably going to be difficult mostly because of heat dissipation issues. You'd probably have to make the case out of metal to act as a big sink for both the Rpi itself and the converter heat outputs.

Also, if you are an Apple user check pout HomeKit SEcure Video support. https://9to5mac.com/2019/11/15/homekit-secure-video-what-to-...

> power over ethernet

This is a non-starter for me since none of the places I would want to put a camera have an Ethernet jack available.

It's not a requirement, just an option. They can also run off of a wall wart or solar panel, with data over WiFi or LTE.

> They can also run off of a wall wart or solar panel

Which is still a nonstarter for me, since none of the locations I would want to put a camera have an electrical jack handy, and many of them don't get enough sunlight to make a solar panel a viable option.

So what do you want? Batteries?

> So what do you want? Batteries?

Yes.

Some of the Arlo gear is battery-powered. I use a couple of their LTE cameras in areas without any other means of connection.

Yes, Arlo also makes good battery operated cameras. They are higher quality than Blink cameras, with higher resolution. On the other hand, Blink cameras have better battery life.

For constant streaming of video you're going to be constantly swapping out batteries as you're powering the camera itself, the ir array, and the wifi.

> For constant streaming of video

I don't need constant streaming of video, only when an event occurs that needs to be recorded.

The camera still has to run for that to work, unless you're also going to power a separate motion sensor and/or a small computer to handle automating the recording; all of this just decreases how long the battery lasts.

> The camera still has to run for that to work, unless you're also going to power a separate motion sensor and/or a small computer to handle automating the recording

Security cameras will have motion sensors built in. Recording would be elsewhere, the camera would just need to stream video for some configurable time if the motion sensor detects an event.

(This is all functionality built into Ring cameras, so for anyone looking for an alternative to them it doesn't seem at all unreasonable to me.)

Eufy (the camera arm of Anker) has battery-powered cameras that seem to be somewhat comparable to Arlo.

See https://www.eufylife.com/products/604/654/battery-camera

Then you don't want a PoE camera. PoE cameras are great if you want to hardwire them somewhere and want to run just one cable.

Not just one cable, but one low power cable. I'd MUCH rather run PoE (no electrician, no permit, no danger, etc.) than running power to a wifi camera.

Blink cameras have excellent battery life, provided there is not frequent movement in the zones you are monitoring.

Exactly what would you consider excellent? Did you measure it personally?

That is another Amazon company so can we note expect the same shenanigans?

Nor did I. But I bought a PoE switch for my garage and installed one near my Front door, two pointing at my driveway, and one pointing at my fence door. PoE is just cat 5 cable, it's considered "low voltage" so you can just run it where you need to. It's much easier than trying to running normal power, which you'd need to do for a wifi connected camera (unless you want to regularly replace batteries).

It's a nice clean mount, I just use the template, drill a 1/2" or so hole for the PoE cable, and then 3 screws to finish the mount. It looks nice and clean, works great, never runs out of power, and is never impacted by the random Wifi issues.

Sure it's a pain to run cat5 everywhere, but I haven't found a better solution, none of my camera locations had power.

Reply on news.ycombinator.com