AI is quickly revolutionizing the security camera industry. Several manufacturers sell cameras which use deep learning to detect cars, people, and other events. These smart cameras are generally expensive though, compared to their “dumb” counterparts. [Martin] was able to bring these detection features to a standard camera with a Raspberry Pi, and a bit of ingenuity.
[Martin’s] goal was to capture events of interest, such as a person on screen, or a car in the driveway. The data for the events would then be published to an MQTT topic, along with some metadata such as confidence level. OpenCV is generally how these pipelines start, but [Martin’s] camera wouldn’t send RTSP images over TCP the way OpenCV requires, only RTSP over UDP. To solve this, Martin captures the video stream with FFmpeg. The deep learning AI magic is handled by the darkflow library, which is itself based upon Google’s Tensorflow.
Martin tested out his recognition system with some cheap Chinese PTZ cameras, and the processing running on a remote Raspberry Pi. So far the results have been good. The system is able to recognize people, animals, and cars pulling in the driveway. His code is available on GitHub if you want to give it a spin yourself!
A home security camera can be great for peace of mind, and keeping an eye on the house while you’re away. The popular option these days is an IP-based device that is accessible over the Internet through an ethernet or wireless connection to your home router. But what if you could cut out the middle man, and instead turn your router itself into the security camera? [Fred] is here to show us how it’s done.
The hack begins by parsing the original router’s firmware. Through a simple text search, a debug page was identified which allowed telnet access to the router to be enabled. This gives access to a root shell, allowing full control over the Linux system running the show.
After backing everything up, [Fred] grabbed the source code from Netgear and recompiled the kernal with USB video and Video4Linux2 support. This allows the router to talk to a standard USB webcam. It’s then a simple matter of using opkg to install software to set up the router to record video when motion is detected.
Overall, it’s fairly straightforward, but [Fred] came up with an ingenious twist. Because the router itself is acting as the security camera, he is able to set up the camera to only arm itself when his smartphone (and thus, [Fred] himself) is not at home. This prevents the recording of footage of [Fred] moving around the house, allowing the router to only record important footage for security purposes.
It’s possible to do great things with routers – most of them are just tiny boxes running Linux anyway. Check out this one used as an online energy meter.
[Mark West] and his wife had a problem, they’d been getting unwanted guests in their garden. Mark’s solution was to come up with a motion activated security camera system that emails him when a human moves in the garden. That’s right, only a human. And to make things more interesting from a technical standpoint, he does much of the processing in the cloud. He sends the cloud a photo with something moving in it, and he’s sent an email only if it has a human in it.
Continue reading “Motion Detecting Camera Recognizes Humans Using The Cloud”
Security in the home — especially a new home — is a primary concern for many. There are many options for security systems on the market, but for those will the skills, taking matters into your own hands can add peace of mind when protected by a system of one’s own design. [Armagan C.] has created their near-ideal multi-sensor security module to keep a watchful eye out for would-be burglars.
Upgrading from their previous Arduino + Ethernet camera — which loved to trigger false alarms — [Armagan] opted for a used Raspberry Pi model B+ camera module and WiFi connection this time around. They also upgraded the unit with a thermal sensor, LPG & CO2 gas sensor, and a motion tracking alarm. [Armagan] has also set up a live streaming feature that records video in 1hr segments — deleting them daily — and circumvented an issue with file descriptor leak by using a crashed drone’s flight controller to route the sensor data via serial port. It is also proving superior to conventional alarms because the custom software negates the need to disarm security zones during midnight trips to the washroom.
Continue reading “Multi Sensor Security Camera Has You Covered”
The Internet of Things is slowly turning into the world’s largest crappy robot, with devices seemingly designed to be insecure, all waiting to be rooted and exploited by anyone with the right know-how. The latest Internet-enabled device to fall is a Motorola Focus 73 outdoor security camera. It’s quite a good camera, save for the software. [Alex Farrant] and [Neil Biggs] found the software was exceptionally terrible and would allow anyone to take control of this camera and install new firmware.
The camera in question is the Motorola Focus 73 outdoor security camera. This camera connects to WiFi, features full pan, tilt, zoom controls, and feeds a live image and movement alerts to a server. Basically, it’s everything you need in a WiFi security camera. Setting up this camera is simple – just press the ‘pair’ button and the camera switches to host mode and sets up an open wireless network. The accompanying Hubble mobile app scans the network for the camera and prompts the user to connect to it. Once the app connects to the camera, the user is asked to select a WiFi connection to the Internet from a list. The app then sends the security key over the open network unencrypted. By this point, just about anyone can see the potential for an exploit here, and since this camera is usually installed outdoors – where anyone can reach it – evidence of idiocy abounds.
Once the camera is on the network, there are a few provisions for firmware upgrades. Usually, firmware upgrades are available by downloading from ‘private’ URLs and sent to the camera with a simple script that passes a URL directly into the shell as root. A few facepalms later, and [Alex] and [Neil] had root access to the camera. The root password was ‘123456’.
While there’s the beginnings of a good Internet of Camera in this product, the design choices for the software are downright stupid. In any event, if you’re looking for a network camera that you own – not a company with a few servers and a custom smartphone app – this would be near the top of the list. It’s a great beginning for some open source camera firmware.
Thanks [Mathieu] for the tip.
Wanting to repair his much-used NES controllers [Michael Moffitt] sourced a replacement for the rubber button pads. They didn’t work all that well but he fixed that by using angle clippers on the part that contacts the PCB traces.
Here’s a neat Claw Game project show-and-tell video. [Thanks David]
We already know that [Bunnie] is building a laptop. Here’s an update on the project.
Hackaday alum [Caleb Kraft] continues his helpful hacking by adding an alternative to clicking an Xbox 360 stick.
[Blackbird] added a camera to the entry door of his house. He didn’t want to forget to shut it off (wasting power) so he built an automatic shutoff.
We’re not really sure what this computational photography project is all about. It takes pictures with the subject illuminated in different colors then combines individual color channels with a MATLAB script.
Finally, [Dave Jones] tears down a Nintendo 64 console on a recent EEVblog episode.
[Zitt] has a security camera which will send him messages any time it detects motion. However annoying this might seem, we’re sure he has his reasons for needing this much immediate feedback. The real problem comes when he goes to view the feed on his iPhone. His solution is to turn the camera’s notifications off, and use his own script to transcode a clip and shoot off an email.
As you can see above, the end result is a concise email that includes the recently captured clip, as well as links to the live feed. He has been storing the clips on an LG N4B2 Network Storage Server (NAS) and since he’s got root access to the Linux system on the device it was an easy starting point for the new system. After he compiled FFmpeg from source (which handles the transcoding) he started work on the script which backs up the recordings and sends the email messages.
One thing he wants to add is a method to clear out the old backup videos. Having encountered a similar issue ourselves we decided to share our one-liner which solves the problem. Find it after the break.
Continue reading “NAS-based transcoding facilitates security cam viewing on iPhone”