By collecting the color, texture, structure and motion feature data of the high-fidelity scene, the device transmits the pedestrian signal detected on the road surface to the spike controller, thereby realizing the blinking prompt of the spike, warning the driver to slow down or stop waiting. At the same time, remind pedestrians to pay attention to the vehicle at the intersection.
Application scenario
It is suitable for major traffic intersections with large traffic flows such as national provincial roads, county and township highways, cities and villages.
Functional characteristics
● High fidelity visual acquisition and processing technology
High-fidelity visual acquisition and processing technology to capture color, texture, structure and motion feature data of the detected scene.
● Feature extraction for different road conditions
Feature classification technology and structure-based motion texture analysis are used to provide positive and negative sample feature extraction for different road conditions.
● Strong anti-interference ability
The video recognition algorithm based on deep learning effectively avoids signal transmission problems caused by weather, environment and equipment interference factors, and reduces cable construction.
● Industrial grade design, reliable and stable
This product has the basic functions of anti-theft, high temperature resistance, cold resistance, heat dissipation and air exhaust. The various types of electrical terminals, overload, leakage and open circuit protection devices, lightning protection devices and other devices are in compliance with the relevant national electrical safety standards.
● Standardized design, strong compatibility and strong scalability
The system adopts an open system architecture and adopts RTP and RSTP standard protocols, which can access relevant data of third-party systems and provide relevant data required by them.
● Easy to operate and easy to maintain
The system is powerful and user friendly. The software design is user-friendly and easy to be mastered, operated and used by ordinary users. The system can basically be in maintenance-free operation, and manual maintenance can be operated remotely.
Video capture | |
sensor type | HD progressive scan CMOS |
Digital noise reduction | Texture noise reduction model |
Wide dynamic range | Digital wide dynamic, ≥120dB |
Focus | Auto focus |
Pedestrian detection | |
Feature extraction | Feature classification technique and structure-based motion texture analysis for positive and negative sample feature extraction of pavement conditions |
Processing technology | High-fidelity visual acquisition and processing technology to capture color, texture, structure and motion signature data of the detected scene |
Recognition algorithm | Video recognition algorithm based on deep learning |
Interface and protocol | |
Interface Protocol | Onvif, GB28181, API |
audio port | 1 input, 1 output |
Video output | 1Vp-pCompositeOutput (75Ω/BNC) |
General specifications | |
Storage function | Support MicroSD/SDHC card storage |
power supply | DC12V/AC24V/POE |
Power consumption | Max16W |
Working temperature / humidity | -10°C--50°C/20%~80%RH |
Size (mm) | 260*139*99 |
weight | 1.3kg |