Antiterrorism
Criminal threat recognition
Automatic system for detection and terrorism of reconnaissance and attack devices, as well as a countermeasure complex.
6 Full-HD IP cameras, 4 sensors, 4 relays, combined logic. High-precision Class M neural networks: weapon (pistol, rifle-carbine, Kalashnikov assault rifle), drone detector, human posture analyzer (falling), person, crowd, car, truck, bicycle-motorcycle, bus, tram/train, face detector.
Gun crime is on the rise, and this is just the beginning.
Speclab has developed a smart video recorder called videoblazer that goes beyond standard security features by detecting any weapon. This process is completely automated, with countermeasures built right in – all in one device.
Videoblazer gives you the ability to react before you even feel a gun pointed at you. Even if your children are home alone and unaware of the danger, the device's built-in power relays can lock doors and connect to the intercom system to warn them before they leave.
Videoblazer's neural network can identify any type of firearm and distinguish between casual carrying and a threat of imminent violence. It works from any angle, with any camera, even those already installed.
Alert systems are integrated into the device, sending messages or sounding alarms. Imagine a school being alerted within seconds. With the increasing number of weapons coming back from the war, it's more likely that they'll find their way into schools. We already have private security firms protecting schools with weapons detection systems, as discussed in another video.
Videoblazer can also monitor weapon handling procedures. Every year, many people die from mistakes in loading and unloading. Now, instructors can be alerted if employees point their weapons at others.
The internet is overflowing with firearms images, allowing us to train our neural network to detect weapons with extremely high accuracy. However, we recognize that cheap Russian cameras have low resolution, despite their popularity. They often have to operate in the dark and with significant interference. No smart camera can handle such demands, but Videoblazer excels with multiple video channels thanks to its cutting-edge neural network chips.
In challenging conditions, we combine the weapon neural network with a "pose" neural network that identifies body parts. This refines threat assessment by confirming the weapon is actually in someone's hands. So, a reflection on a camera that looks like a gun or a bird on a fence in bad weather won't trigger false alarms.
Beyond weapons, Videoblazer detects various threats like drones, faces, license plates, fights, fence climbing, falls, even QR codes, clothing styles, money, and credit cards.
We've been supplying this equipment to our partners for years.
Chinese technology is no match. Their so-called smart systems often have a tolerance for a certain number of false alarms. But the threshold is usually very high, leading operators to ignore the alerts or disable them entirely – depending on their temperament. Videoblazer surpasses both Chinese programs and European computers in this regard. Plus, it's affordable. This small, weatherproof box requires no special space, cooling, or maintenance.
Shooter’s stance