BENTENG-7 face recognition and access control all-in-one machine Introduction: BENTENG-7 is a deep learning based facial front-end recognition tablet (DV300 scheme) with very flexible features. It adopts the HiSilicon industrial embedded Linux platform, built-in neural network computing processor, and a 2 megapixel fully high-definition WDR color camera to easily cope with backlit, dark, and sunny environments. It can store 20000 face libraries locally for offline recognition and comparison. The device has rich tail line interfaces and flexible configurations, Optional QR code recognition including health QR code, external card swiping, second-generation ID card, etc. characteristic: 1. Low illumination CMOS, 0.01Lux @ (F1.2, AGC ON); 2. High performance embedded processor, integrating image acquisition, facial detection, facial tracking, facial comparison, and live body judgment; 3. Autonomous CNN neural network algorithm, dynamic recognition without special cooperation; 4. Support panorama and local close-up image output; 5. Support identification of portraits, safety helmets, masks, etc; 6. Support QR code recognition; 7. Support the person certificate comparison mode; 8. Recognition speed ≤ 80ms, recognition rate ≥ 99%; 9. Support 2W portrait library; 10. Support video overlay; Supports the Wigan protocol; 11. Lightning and surge protection measures; Technical parameters: Sensor: 1/2.8 "progressive scan CMOS Minimum illumination: 0.01Lux @ (F1.2, AGC ON) Shutter: 1/100s~1/10000s Lens: Dual Sony 6mm lens Day and night parameters: adaptive Wide dynamic: support Voice broadcasting: support, support TTS voice broadcasting customization; Video standard: H.264 Bit rate: 1024Kbps~4Mbps Video size: 1920 (H) * 1080 (V) Frame rate: 25fps Video settings: exposure (shutter), gain, contrast, saturation, face exposure compensation Working mode: online, offline, automatic Identification content: portrait, optional (QR code recognition, advertising playback, card swiping (IC+ID)) Identification type: 1: N Detection type: In vivo detection Recognition speed: ≤ 80ms Recognition rate: ≥ 99% Recognition distance: 0.5-2 meters Portrait database: 20000 sheets Portrait database import: single image, batch image, real-time captured image import, platform distribution Image format: JPEG encoding Capture results: panorama, local close-up Picture size: 1920 * 1080 (panorama), local close-up according to the actual picture proportion Local storage: Built-in 8GB eMMC for network interruption and continuous transmission Snap storage: close-up image: single machine supports 640000 copies (insert 64G TF card) Panorama: a single machine supports 320000 pieces (insert 64G TF card) Output mode: Wigan 26/34/66, relay Trigger method: video trigger, card swiping, ID swiping, QR code swiping Secondary development: Multi platform (Linux, Arm, Windows), multi language (Java, C++, C #) Protocol support: ONVIF, TCP/IP, HTTP, FTP, 485, GB28181, Wigan, DNS, NTP General functions: heartbeat, password protection, NTP timing Power supply: DC 12V 2A Power consumption: ≤ 8W Working temperature: -30 ℃~+70 ℃ Working humidity: 20% -90% Waterproof grade: IP66 Protection measures: lightning protection and surge prevention
Introduction: BENTENG-7 is a deep learning based facial front-end recognition tablet (DV300 scheme) with very flexible features. It adopts the HiSilicon industrial embedded Linux platform, built-in neural network computing processor, and a 2 megapixel fully high-definition WDR color camera to easily cope with backlit, dark, and sunny environments. It can store 20000 face libraries locally for offline recognition and comparison. The device has rich tail line interfaces and flexible configurations, Optional QR code recognition including health QR code, external card swiping, second-generation ID card, etc.
