It‘s still a bit flaky getting video acceleration (not talking about object detection but video decoding) working but after that it is one of the best solutions for live object detection I‘ve ever tried: no more small animals waking me up in the night.
P.S.: I‘m also supporting them with a yearly? subsciption to train the „A.I.“ model against false positives I provide which increased the accuracy even more.
As an alternative, you might also want to check out scrypted which offers a lot of cross-integration features and hardware optimized local AI processing (eg on MacMinis M*). Developer is super responsive in the discord.
I use Ubiquiti Protect Cameras and recently bought a AI key[1] which adds license plate and facial recognition features to all cameras even non-AI enabled models. It works really well and of course all 100% self-hosted.
OpenIPC is an alternative open firmware for your IP camera.
OpenIPC is an open source operating system from the open community targeting for IP cameras with ARM and MIPS processors from several manufacturers in order to replace that closed, opaque, insecure, often abandoned and unsupported firmware pre-installed by a vendor.
Frigate has really done a fantastic job packing everything together.
For basic needs go2rtc [0] or MediaMTX [1] can be enough.
But once you need some form of intelligence on top AFAIK unfortunately there is no unixiy tool that can take a stream and easily define and apply a model on it. You will have to code something in python.
My step brother has been asking me to help him setup a load of cameras for watching his marron ponds. he has foxes, crows and humans stealing from his ponds.
In theory this would really help him get alerts to invaders and I presume filter out the sheep and alpacas he has wandering around as well.
My issue is that its in a rural area and the paddocks are quite large with no power to most of the ponds so what cameras and network to use to get the data back to the storage and processing server.
Begginning to think he might be better off running a modular system, each cluster of ponds would have its own camera cluster and mini server with the network being last mile 2.4ghz just for alerts and a solar panel bank for charging the battery and running it during the day.
What would I get away with here? N100 mini device? processing maybe 6 cameras?
I've been running Frigate for many years, using a PN50 NUC and a Coral USB dongle, the Coral is a must, at least in my case. I had a full blown Ubiquiti/Unifi setup with cameras + their software. Way to many false alarms compared to Frigate. Now I run 10+ cameras with 24/7 recording and alarms with images pushed to Telegram. The identification is instant as well as the telegram message.
Running a mix of Ubiquti/TP-Link VIGI+TAPO/Reolink. I'm running everything in containers and everything works perfect!
Where would I start if I wanted to do stuff on a video, but not necessarily live? Like, say I have a 5h video and want to extract the frames of each car passing when it's at a certain spot, for instance. Or all of those with a driver holding a phone or whatever. Are there good frameworks for this, or would I have to split the video into a million frames and run something on each one?
I've been running Frigate for more than two years now and it beats the hell out of any system I've tried in terms of detection speed and reliability. For context, I've tried Ring, Tapo cameras, and also Eufy security. Today I have turned away from all the cameras except for the Tapo cameras now serving RTSP streams into my Frigate instance. I have also blocked them from accessing the internet and that gave it complete privacy by default.
Eufy Security started showing advertisements about their new products whenever I tap on a motion detected notification. They prioritize their ads over your own security which is ridiculous. Not just that, some of their clips stored in their cloud storage would never open despite the fact I used to pay them my membership fees every month. They were also caught storing passwords and other security credentials in plain text. Thanks to them, they were the primary motivation for me to move away from using those proprietary platforms and look for something self-hosted.
I got Frigate running on my old hardware with hardware acceleration enabled via RX 550 GPU and detection is always under one second. I wrote a small app that uses Frigate API to grab screenshots and send me notifications via Telegram and Pushover. It's been very self-sustainable for two years now. I only had to restart the service two times in all of this time. I am also using some tunneling from my VPS into the locally hosted Frigate running on my home server and it's just been flawless. Thanks to this amazing project.
Why are people still installing security cameras that are monitored by them? They increase stress level and felt insecurity. They do not make you feel secure, say psychological studies. You probably think more about burglaries and dead spaces in your setup and actively monitor for these in your daily lives, where for 99.8 % of people this should be a non-topic.
If you want to install them for later police work, that still seems tedious and you might require off-site backup. In public places we often have CCTV of people, but unless you have number signs on vehicles, they seem to not help with conviction rates by much.
In my previous apartment, the landlady had zero sense of decency and would let herself in to snoop around.
I use these devices because I can factually know that nobody has entered my home while I was gone. It is peace of mind. I don't think about burglaries or whatever. I think about how my landlord or a property manager or rotating cast of anonymous maintenance people have a key and the only reason they don't abuse it is because of decency.
I get notifications when packages are delivered which limits the window for porch pirates.
I have a daily news feed of animal activity so I can see what the little neighborhood cats, raccoons, and skunks were up to last night. I was originally using it to alert me when the neighborhood stray was on the back porch so I could come down and feed her (without risking other critters finding the food)
I am installing a doorbell one this week. I got a package delivered monday according to the tracker but its not here. It would be nice to have had the camera already so I could see if someone took the package or if its still potentially not yet delivered. Neighbors have gotten packages stolen plenty so it is a real risk.
What is the cheapest way to do something like this in a DIY way with FOSS? Assuming you have to buy the computer, and any other hardware. Assuming also near real-time processing and reasonably high accuracy.
I run Frigate with 5 IP cameras (3 Hikvisions, 2 Amcrests) and 1 USB camera. I'm using a USB Coral TPU, which does a good enough job that Frigate can keep up with an average of only 30% CPU usage on an old Dell with 4 core i7-6700.
Frigate's better than anything else I tried, but not perfect. As mentioned in another thread, it has some issues with codecs from some cameras (playing clips from Amcrests is fine, Hikvisions not so much) and therefore you may need to transcode. Also it has no built in option for sending your recorded clips offsite; theoretically you could mirror its storage directory, but as far as I've found it's not organized in a way that you can separate just important events.
Side note. I’m surprised we’re not doing more with LLMs as far as image and video processing. We now have some level of imaging understanding in a box (and some common sense). Seems like there would be millions of possibilities.
Is manufacturing using it for anything? More security applications?
43 comments
[ 1.3 ms ] story [ 55.6 ms ] threadP.S.: I‘m also supporting them with a yearly? subsciption to train the „A.I.“ model against false positives I provide which increased the accuracy even more.
[1] https://store.ui.com/us/en/category/all-cameras-nvrs/product...
It has a very nice integration with homeassistant.
https://openipc.org/?locale=en
For basic needs go2rtc [0] or MediaMTX [1] can be enough. But once you need some form of intelligence on top AFAIK unfortunately there is no unixiy tool that can take a stream and easily define and apply a model on it. You will have to code something in python.
- [0] https://github.com/AlexxIT/go2rtc
- [1] https://github.com/bluenviron/mediamtx
In theory this would really help him get alerts to invaders and I presume filter out the sheep and alpacas he has wandering around as well.
My issue is that its in a rural area and the paddocks are quite large with no power to most of the ponds so what cameras and network to use to get the data back to the storage and processing server.
Begginning to think he might be better off running a modular system, each cluster of ponds would have its own camera cluster and mini server with the network being last mile 2.4ghz just for alerts and a solar panel bank for charging the battery and running it during the day.
What would I get away with here? N100 mini device? processing maybe 6 cameras?
If you want to start just remember to avoid h.265 cameras so you don't need to transcode since few clients and browsers support it.
Running a mix of Ubiquti/TP-Link VIGI+TAPO/Reolink. I'm running everything in containers and everything works perfect!
Eufy Security started showing advertisements about their new products whenever I tap on a motion detected notification. They prioritize their ads over your own security which is ridiculous. Not just that, some of their clips stored in their cloud storage would never open despite the fact I used to pay them my membership fees every month. They were also caught storing passwords and other security credentials in plain text. Thanks to them, they were the primary motivation for me to move away from using those proprietary platforms and look for something self-hosted.
I got Frigate running on my old hardware with hardware acceleration enabled via RX 550 GPU and detection is always under one second. I wrote a small app that uses Frigate API to grab screenshots and send me notifications via Telegram and Pushover. It's been very self-sustainable for two years now. I only had to restart the service two times in all of this time. I am also using some tunneling from my VPS into the locally hosted Frigate running on my home server and it's just been flawless. Thanks to this amazing project.
Why are people still installing security cameras that are monitored by them? They increase stress level and felt insecurity. They do not make you feel secure, say psychological studies. You probably think more about burglaries and dead spaces in your setup and actively monitor for these in your daily lives, where for 99.8 % of people this should be a non-topic.
If you want to install them for later police work, that still seems tedious and you might require off-site backup. In public places we often have CCTV of people, but unless you have number signs on vehicles, they seem to not help with conviction rates by much.
I use these devices because I can factually know that nobody has entered my home while I was gone. It is peace of mind. I don't think about burglaries or whatever. I think about how my landlord or a property manager or rotating cast of anonymous maintenance people have a key and the only reason they don't abuse it is because of decency.
I have a daily news feed of animal activity so I can see what the little neighborhood cats, raccoons, and skunks were up to last night. I was originally using it to alert me when the neighborhood stray was on the back porch so I could come down and feed her (without risking other critters finding the food)
Frigate's better than anything else I tried, but not perfect. As mentioned in another thread, it has some issues with codecs from some cameras (playing clips from Amcrests is fine, Hikvisions not so much) and therefore you may need to transcode. Also it has no built in option for sending your recorded clips offsite; theoretically you could mirror its storage directory, but as far as I've found it's not organized in a way that you can separate just important events.
They use the abbreviation NVR in the first sentence without saying what it means.
It means "networked video recorder".
Please don't do that. Not everyone who comes across your site is a member of your particular niche.
Unfortunately, the USB Accelerator is very hard to buy even at 3x retail.
Is manufacturing using it for anything? More security applications?