X3221-C 2MP Ultra-Starlight Face Recognition Infrared Dome Camera
The Huawei X3221-C is an intelligent camera designed for facial recognition at traffic checkpoints. Its face recognition is primarily based on computer vision, integrating image processing and deep learning technologies to extract facial feature values from massive amounts of video data and perform comparative analysis, building an accurate, efficient, and stable facial recognition system for customers.
Face Recognition: Built-in face recognition algorithm supports face detection and tracking in video, and automatic recognition of detected faces. A single camera can hold up to 100,000 faces.
The camera automatically recognizes scene features and adjusts imaging parameters in real time to ensure image quality under different lighting conditions, foggy weather, and varying movement speeds.
Software-Defined: Based on powerful chip computing power and an open software architecture, the camera supports online loading of new algorithms and independent algorithm upgrades, without interrupting service during upgrades and loading.
Technical Specifications
Image Sensor: 1/1.8" 2MP progressive scan CMOS
Effective Image Size: 1920(H) * 1080(V)
Minimum Illumination: Color: 0.0004 Lux @(F1.2, AGC ON)

Monochrome: 0.0001 Lux @(F1.2, AGC ON) ON) Wide dynamic range 120dB White balance Auto/Auto tracking/Manual Strong light suppression 

supported Electronic defogging Auto/Manual/Off Digital noise reduction

 supported Intelligent adaptive 2D/3D noise reduction 

Lens & Lighting Focal length 8-32mm Maximum aperture F1.6 Aperture type P-Iris Lighting distance 15m infrared

 Video compression format H.265/H.264/MJPEG Maximum resolution 1920×1080 (1080P) 

Frame rate 60Hz: 60/30/20~1fps; 50Hz: 50/25/20~1fps Multi-stream five-stream,

supports 1080p dual-channel independent high-definition encoding

1/1.8" progressive scan CMOS image sensor
Built-in NPU neural network engine, greatly improving the
performance of deep learning algorithms
Facial recognition at the front end, supporting the alarm
output based on the object match with 100,000 blacklist records
Simultaneous snapshot of the face, person, and head and shoulder
Simultaneous detection and tracking of 100 targets
Face capture supported for faces at an ultra-large deviation angle
Face tracking, filtering, and evaluation to generate the
optimal face snapshot and reduce repeated snapshots
Facial attribute recognition and personal attribute recognition
AI-based scene adaptation. The camera can detect scenarios
in real time and make targeted setting optimization to improve imaging quality.
Software-defined camera, supporting online loading and upgrade of algorithms
Extra265 intelligent encoding, reducing 70% of the requirements on storage and bandwidth resources
Area cropping, viewing fine details through limited bandwidth 
Stream smoothing, meeting various requirements for video quality and fluency
Traffic shaping, enabling precise control over instantaneous
burst of video streams and delivering smooth video without packet loss
TCP acceleration, enabling the network to carry video streams of higher quality