Community CCTV Project

[Bl4d3]

Hey all
Wanting to get some input on this idea.
Our area has had a massive increase in crime in the last 2 years. Hijackings, muggings, break-ins, people being held at gun point in their home, a couple of kidnappings, and even an alleged s****l assault in a public park. Sounds like the wild west, but its actually a suburb in Bellville, Cape Town.

Our NHW is obviously trying to do their best in conjunction with local armed response companies considering the limited support they receive from government, however it seems to not be enough.

With the limited funds the NHW receives in donations at through the POCS R10000 it receives once a year, we're looking at setting up a community CCTV project, not only to monitor the high crime spots, but, to get those who may not be able to physically patrol to virtually patrol and monitor cameras for an hour. Yes they would have motion detection but it is a way to get live involvement.

Individual camera points with 2 x POE cameras, POE switch and a NVR at each location would be rather expensive, quickly nearing R3000 - R4000 before installation.

With that, I had an idea.
I've come into quite a few cameras that I can potentially get at work for next to nothing. They are 2MP cameras but more than enough for an initial roll out.
These 2 cameras would connect to a POE switch, which would in turn connect to a router such as a Cudy TR1200 for the purpose of using WireGuard to connect to my home server that has Frigate installed.

Now I've tested this and this has worked well, however, I hit a snag, and thats data usage.
Running 18 x 2MP cameras would likely generate roughly 14TB of data used per month.

Now this isnt anywhere near those on the datahogs list, but its still up there.

Has anyone had experience with something like this and if this would be problematic?
I have also popped my ISP an email with the proposition.

 

[Mike Angelo]

Maybe contact Vumacam. As far as I know they are not operating in the Cape so possibly they would assist with advice.

 

[PretoriaWISP]

Well local network for cameras would be rather crazy. I can say that internet traffic for a restaurant on our network has about 25-35 cameras and their internet traffic for the month is about 1.1TB upload. If you plan to do this I'd also recommend contacting ICASA for any sort of exempt licensing that could be applicable if the plan is to run some sort of LAN network for all of them. Send me a pm happy to give you guys some guidance

 

[Bl4d3]

Well local network for cameras would be rather crazy. I can say that internet traffic for a restaurant on our network has about 25-35 cameras and their internet traffic for the month is about 1.1TB upload. If you plan to do this I'd also recommend contacting ICASA for any sort of exempt licensing that could be applicable if the plan is to run some sort of LAN network for all of them. Send me a pm happy to give you guys some guidance

Traffic for a local network would be relatively low as that would be recording with in a local network and not touching the internet, unless the owner is streaming the substream 24/7 to off site, which would explain the 1TB of traffic.

We looked at setting up line of sight bridges but that is incredibly expensive. Unfortunately the NHW is severely underfunded and operates almost entirely off of donations and goodwill, while trying to fulfil the role of community safety while not really receiving the support needed from government, other than a yearly R10000 from POCS which is applied for.

I sent a mail off to Atomic to hear their thoughts as well and if they would allow such high traffic. I see most ISP's dont have a limit, but I thought I'd rather ask before we invest in each point

 

[PretoriaWISP]

Traffic for a local network would be relatively low as that would be recording with in a local network and not touching the internet, unless the owner is streaming the substream 24/7 to off site, which would explain the 1TB of traffic.

We looked at setting up line of sight bridges but that is incredibly expensive. Unfortunately the NHW is severely underfunded and operates almost entirely off of donations and goodwill, while trying to fulfil the role of community safety while not really receiving the support needed from government, other than a yearly R10000 from POCS which is applied for.

I sent a mail off to Atomic to hear their thoughts as well and if they would allow such high traffic. I see most ISP's dont have a limit, but I thought I'd rather ask before we invest in each point

You know what, I'd just subscribe to whoever just be open with the ISP then about the intent they may ask for a business account all depends, do you have fibre in the area?. @PBCool maybe you can shed light on the setup for something like this on WAN

 

[PBCool]

You know what, I'd just subscribe to whoever just be open with the ISP then about the intent they may ask for a business account all depends, do you have fibre in the area?. @PBCool maybe you can shed light on the setup for something like this on WAN

Hi, I think there are a few customers that do this on our network.

The major question would also be how many lines would you commission? Or is it a single location with lots of cameras?

You also obviously run the risk of your house losing power or fibre breaking and the whole thing collapses with it.

 

[Bl4d3]

Hi, I think there are a few customers that do this on our network.

The major question would also be how many lines would you commission? Or is it a single location with lots of cameras?

You also obviously run the risk of your house losing power or fibre breaking and the whole thing collapses with it.

Fibre failing would be the systems achilles heel, as there would be no way to have a fallback to LTE due to the amount of data.
Each camera point would consists of 2 cameras and would be on a residents fibre, assuming they'd be happy to.
Considering these would be relatively low, I cant see that being the issue, rather the 12 - 14TB of data coming in on my line, which can easily handle it (1000/250).

This would be a phase 1 of the wider camera initiative to try and get more community involvement and funding.

At a later stage, the system could be redesigned or upgraded to rather have NVR's at each point to record to, which would strengthen the system considerably. But, this would only happen after more of the community becomes involved, which we believe would be possible if authorised and vetted community members are able to rather do a camera patrol vs driving around the area and putting ones self in harms way

 

[PBCool]

Fibre failing would be the systems achilles heel, as there would be no way to have a fallback to LTE due to the amount of data.
Each camera point would consists of 2 cameras and would be on a residents fibre, assuming they'd be happy to.
Considering these would be relatively low, I cant see that being the issue, rather the 12 - 14TB of data coming in on my line, which can easily handle it (1000/250).

This would be a phase 1 of the wider camera initiative to try and get more community involvement and funding.

At a later stage, the system could be redesigned or upgraded to rather have NVR's at each point to record to, which would strengthen the system considerably. But, this would only happen after more of the community becomes involved, which we believe would be possible if authorised and vetted community members are able to rather do a camera patrol vs driving around the area and putting ones self in harms way

Yeah so a few caveats there, you are looking to use random residents fibre services, firstly if they have, what service is it. Is it CGNATed, then managing tunnels back to your house. When there is an issue who and how do you go about troubleshooting that. It may be in breach of the multiple ISP terms hosting cameras.

The risks involved with all of that alone.

I understand where you are coming from, trying to make an effective security system with very little resources. But in most cases of leveraging communities its a nightmare to manage .

 

[Bl4d3]

Yeah so a few caveats there, you are looking to use random residents fibre services, firstly if they have, what service is it. Is it CGNATed, then managing tunnels back to your house. When there is an issue who and how do you go about troubleshooting that. It may be in breach of the multiple ISP terms hosting cameras.

The risks involved with all of that alone.

I understand where you are coming from, trying to make an effective security system with very little resources. But in most cases of leveraging communities its a nightmare to manage .

Very valid points. I will need to do some digging and reading

 

[horticulturist]

Where is CTWUG when you need it!

Great to see you taking the initiative on this community project. I immediately thought of setting up a WUG (Wireless User Group) for this purpose. CTWUG was a legendary city wide project that did exceptionally well at the time., a huge portion of the city was connected. Scaling this down to a residential suburb level might be a viable option for a local CCTV network.

Disclaimer: This bypasses the need for a utilizing an for an internet connection through and ISP but I have no idea how viable this idea would be (particularly financially speaking), just throwing it out there.

(I had Gemini lay out the following):

1. The Local CCTV Network Layout
Instead of routing outward to the web, all network paths will converge on a Central Recording Hub located at a secure, trusted location within the neighborhood.

[Camera Node 1] \

===> (Wireless Backbone Links) ===> [Central Recording Hub] (NVR + Storage + Monitor)​

[Camera Node 2] /

• The Camera Nodes: Cameras are mounted on houses or brackets facing the street. Each camera plugs directly into that home's outdoor wireless antenna (CPE).
• The Central Hub: One central home hosts the Network Video Recorder (NVR). A powerful multi-directional antenna (Sector Antenna) is mounted on this roof to pull in the wireless video feeds from all the surrounding camera nodes simultaneously.

2. Tailored Hardware Requirements for Video Traffic

Because continuous video streaming requires constant, stable bandwidth, choosing the right wireless frequency and camera setup is critical to avoid stuttering or dropped frames.

Component Choice & Best Practice Why It Matters Approx. Cost (ZAR)

Wireless Frequency: 5 GHz Outdoor Antennas (e.g., TP-Link CPE510 or Ubiquiti LiteBeam) The standard 2.4 GHz frequency is too crowded by home routers. 5 GHz offers the clear, high-speed highway needed for multiple video streams.R900 – R1,300 per node

The Cameras: IP Cameras (PoE) with H.265 compression IP cameras send digital data over network cables. The H.265 compression format shrinks video file sizes by 50% without losing quality, saving wireless bandwidth. R600 – R1,200 per camera

Central Recorder: Network Video Recorder (NVR) with a Surveillance Hard Drive. An NVR aggregates all IP streams into one place. Hard drives rated for surveillance (like Western Digital Purple) are designed to handle 24/7 continuous writing. R2,500 – R4,000 (Hub total)

Local Viewing: Old HDMI Monitor or TV plugged directly into the NVR. Since there is no internet to view feeds on mobile phones away from home, this monitor acts as the local "control room" for the neighborhood watch.

 

[Bl4d3]

Where is CTWUG when you need it!

Great to see you taking the initiative on this community project. I immediately thought of setting up a WUG (Wireless User Group) for this purpose. CTWUG was a legendary city wide project that did exceptionally well at the time., a huge portion of the city was connected. Scaling this down to a residential suburb level might be a viable option for a local CCTV network.

Disclaimer: This bypasses the need for a utilizing an for an internet connection through and ISP but I have no idea how viable this idea would be (particularly financially speaking), just throwing it out there.

(I had Gemini lay out the following):

1. The Local CCTV Network Layout
Instead of routing outward to the web, all network paths will converge on a Central Recording Hub located at a secure, trusted location within the neighborhood.

[Camera Node 1] \

===> (Wireless Backbone Links) ===> [Central Recording Hub] (NVR + Storage + Monitor)​

[Camera Node 2] /

• The Camera Nodes: Cameras are mounted on houses or brackets facing the street. Each camera plugs directly into that home's outdoor wireless antenna (CPE).
• The Central Hub: One central home hosts the Network Video Recorder (NVR). A powerful multi-directional antenna (Sector Antenna) is mounted on this roof to pull in the wireless video feeds from all the surrounding camera nodes simultaneously.

2. Tailored Hardware Requirements for Video Traffic

Because continuous video streaming requires constant, stable bandwidth, choosing the right wireless frequency and camera setup is critical to avoid stuttering or dropped frames.

Component Choice & Best Practice Why It Matters Approx. Cost (ZAR)

Wireless Frequency: 5 GHz Outdoor Antennas (e.g., TP-Link CPE510 or Ubiquiti LiteBeam) The standard 2.4 GHz frequency is too crowded by home routers. 5 GHz offers the clear, high-speed highway needed for multiple video streams.R900 – R1,300 per node

The Cameras: IP Cameras (PoE) with H.265 compression IP cameras send digital data over network cables. The H.265 compression format shrinks video file sizes by 50% without losing quality, saving wireless bandwidth. R600 – R1,200 per camera

Central Recorder: Network Video Recorder (NVR) with a Surveillance Hard Drive. An NVR aggregates all IP streams into one place. Hard drives rated for surveillance (like Western Digital Purple) are designed to handle 24/7 continuous writing. R2,500 – R4,000 (Hub total)

Local Viewing: Old HDMI Monitor or TV plugged directly into the NVR. Since there is no internet to view feeds on mobile phones away from home, this monitor acts as the local "control room" for the neighborhood watch.

We considered this, and this is one of the 3 ideal ways to set up a community CCTV network.
In fact, the cost for this is similar to having a NVR at each point.

We have a high point in the centre of the area which makes this even more viable, with all points having direct line of site to it.
In an ideal world, we'd do either of the 2.

The only reason I've considered this route is due to cost. It would be considerably cheaper to get a network set up almost immediately, in which the community can then slowly transition over to a NVR per point or community network.
I estimate around R1000 per point

 

[rustypup]

Raspberry Pi with Infrared cam and tiny LLM for image processing/event generation. Cheap as chips, aside from Solar panel and securing said panel.

Basic data package and you're good to go.

 

[Sparlettac]

We considered this, and this is one of the 3 ideal ways to set up a community CCTV network.
In fact, the cost for this is similar to having a NVR at each point.

We have a high point in the centre of the area which makes this even more viable, with all points having direct line of site to it.
In an ideal world, we'd do either of the 2.

The only reason I've considered this route is due to cost. It would be considerably cheaper to get a network set up almost immediately, in which the community can then slowly transition over to a NVR per point or community network.
I estimate around R1000 per point

What's the other 2 ways from the 3 ideal ways? Just out of interest

 

[Bl4d3]

What's the other 2 ways from the 3 ideal ways? Just out of interest

1. Community network using line of sights dishes / antennas
2. NVR per point
3. Using cameras that can take high endurance SD cards such as the Vigi C340.

Point 3 is fairly promising as Vigi's system can get setup as sites, granting multiple access for viewing, while recording is done on site. The main downside is you lose the recording should someone steal the camera. But, cheaper than setting up a NVR for 2 cameras.

 

[PretoriaWISP]

cptwug is still alive the last I checked, most those guys now do Meshtastic (maybe those guys could give you a better way with security) our wug in pta is dead. My company being formed from one of the admins and providers of said network. Anycase it would be far more simple doing this via wireless but again you'll have to reachout to icasa for this or do a hybrid fibre wireless network mesh it. If price is a worry you can do mikrotik or even reyee 460 dishes or even ubnt if push comes to shove on managing the network better.

 

[CrypticZA]

Fibre failing would be the systems achilles heel, as there would be no way to have a fallback to LTE due to the amount of data.
Each camera point would consists of 2 cameras and would be on a residents fibre, assuming they'd be happy to.
Considering these would be relatively low, I cant see that being the issue, rather the 12 - 14TB of data coming in on my line, which can easily handle it (1000/250).

This would be a phase 1 of the wider camera initiative to try and get more community involvement and funding.

At a later stage, the system could be redesigned or upgraded to rather have NVR's at each point to record to, which would strengthen the system considerably. But, this would only happen after more of the community becomes involved, which we believe would be possible if authorised and vetted community members are able to rather do a camera patrol vs driving around the area and putting ones self in harms way

I work for a company with about 150 - 200 cameras in our warehouse and addistional AI camera's all around the property etc.

Our data usage per month with 250 staff memebers working minimum 5 days a week is about 9TB. Camera quality is also super clear

Current usage for the month

 

[deweyzeph]

We successfully got a community CCTV project up and running where I live in Cape Town. The hardest part was convincing residents to pay for it, but we had community meetings, a professional business plan was put together and we managed to raise close to about R1 000 000 to get it up and running (along with ongoing annual costs). We got a professional CCTV company to set the cameras up all around our neighbourhood including 2 x NPR cameras linked to the SAPS crime intelligence network at the 2 entrances to our neighbourhood and we also pay for a security company to monitor the cameras 24/7 with special focus on the hours of darkness. Crime in our area has dropped to almost zero since the cameras were installed. Once it became obvious that the cameras caused a substantial drop in crime it became easier and easier to fundraise every year to keep the project going.

Good luck, all I can say is that without adequate funding and an ongoing commitment from residents to keep funding the project, you will struggle to get this off the ground. Good cameras aren't cheap and monitoring them isn't cheap either. But it's undeniable that a good CCTV network in area that is actively monitored can completely turn a crime situation around in an area.

Edit: just wanted to say as well that just having cameras alone is not enough. They need to be actively monitored and there needs to be consequences for people who are caught acting suspiciously. In our case the security company monitoring them has a patrol car in the neighbourhood at night that is available to respond immediately to any potential or ongoing incidents caught on camera.

 

[Bl4d3]

We successfully got a community CCTV project up and running where I live in Cape Town. The hardest part was convincing residents to pay for it, but we had community meetings, a professional business plan was put together and we managed to raise close to about R1 000 000 to get it up and running (along with ongoing annual costs). We got a professional CCTV company to set the cameras up all around our neighbourhood including 2 x NPR cameras linked to the SAPS crime intelligence network at the 2 entrances to our neighbourhood and we also pay for a security company to monitor the cameras 24/7 with special focus on the hours of darkness. Crime in our area has dropped to almost zero since the cameras were installed. Once it became obvious that the cameras caused a substantial drop in crime it became easier and easier to fundraise every year to keep the project going.

Good luck, all I can say is that without adequate funding and an ongoing commitment from residents to keep funding the project, you will struggle to get this off the ground. Good cameras aren't cheap and monitoring them isn't cheap either. But it's undeniable that a good CCTV network in area that is actively monitored can completely turn a crime situation around in an area.

Reminds me of when we ran the Ridgeworth CCTV project. We also managed to raise near R1million and it worked.

LPR tech and monitoring has become drastically cheaper. The previous NHW Exco managed to get funds for 2 points, totalling R33000 with perpetual monitoring from SJC. There is just 1 massive caveat with LPR, and that is that there has to be a number plate on the car linked to a crime for it to work.

Its great for tracking a stolen or hijacked car, but doesnt really work for tracking those who've come in to commit a crime.

The type of crime our area also faces is more done by those on foot. Because of our proximity to Strand Road and Bill Bezuidenhout, we're getting a lot of the gangsters in thats squatting under Tienie Meyer bypass near Bellville Station.

 

[deweyzeph]

Reminds me of when we ran the Ridgeworth CCTV project. We also managed to raise near R1million and it worked.

LPR tech and monitoring has become drastically cheaper. The previous NHW Exco managed to get funds for 2 points, totalling R33000 with perpetual monitoring from SJC. There is just 1 massive caveat with LPR, and that is that there has to be a number plate on the car linked to a crime for it to work.

Its great for tracking a stolen or hijacked car, but doesnt really work for tracking those who've come in to commit a crime.

The type of crime our area also faces is more done by those on foot. Because of our proximity to Strand Road and Bill Bezuidenhout, we're getting a lot of the gangsters in thats squatting under Tienie Meyer bypass near Bellville Station.

You'd be surprised how many dodgy people driving vehicles with flagged numberplates drive through our area. There's only 2 entrances to our area and it's on a major route so plenty of dumb criminals drive through and don't even know they've been clocked.

 

[B-1]

You dont have to run 60FPS on these even 5FPS is fine with local recording as evidence if something happens.

I agree with the high site idea then just do low frame rate and you are covered for the most part.

The bigger problem is the monitoring, it needs to be an early warning system so your CPF or security can be alerted.