The embedded deep learning algorithm uses massive pictures and video resources as the roadbed, and extracts the target features through the machine itself to form a deep face image for learning, which greatly improves the detection rate of the target face.
The embedded deep learning algorithm uses massive pictures and video resources as the roadbed, and extracts the target features through the machine itself to form a deep face image for learning, which greatly improves the detection rate of the target face.
Support 5 kinds of intelligent resource mode switching: face capture, mixed target detection (default), road monitoring, smart event, smart city management
Face capture: Supports detection, tracking, capture, scoring, and screening of moving faces, outputting the best face capture, up to 60 simultaneous detections per frame, support for front-end face comparison, and support for up to 3 face databases management, supports up to 90,000 face databases
Hybrid target detection: support face + human body capture + attribute extraction, track and score targets, output optimal snapshot images, and detect up to 60/frame at the same time
Road monitoring: support for back-to-back vehicle snapshots, license plate, sub-brand, body color, and vehicle type. Support mixed traffic detection of pedestrians, non-motor vehicles and motor vehicles
Smart events: support out-of-bounds detection, area intrusion detection, entry area detection, leaving area detection, loitering detection, people gathering detection, fast motion detection, parking detection, item left detection, item retrieval detection Measurement
Smart city management: support 1-24 hours detection cycle, support the detection of illegal stalls
Support low bit rate, low latency, ROI enhanced coding, SVC adaptive coding technology
Support fog penetration, electronic image stabilization
Support WDR 120dB
Supports the highest resolution of 12 million pixels
Support five-stream technology, dual-channel high-definition, support simultaneous 20-channel streaming
Power supply: AC24V: 24V±20%, waterproof grade IP67
Model: STB-AIC342
Camera
Sensor Type1/1.8” Progressive Scan CMOS
Minimum IlluminationColor: 0.0005Lux @ (F1.2, AGC ON) B&W: 0.0001 Lux @ (F1.2, AGC ON), 0 Lux with IR
Shutter1 second to 1/100,000 second
Day and night switching modeICR infrared filter type
WDR Ultra-wide dynamic range up to 120dB, monitoring in indoor backlight environment
Lens
Focal Length & Field of View
(Zoom) 2.8-12mm @ F1.2, horizontal field of view: 107.4°~39.8°, vertical field of view: 56.1°~22.4°, diagonal field of view: 130.0°~45.7°
(Zoom) 8-32mm @ F1.6, horizontal field of view: 37.7°~15.1°, vertical field of view: 21.0°~8.7°, diagonal field of view: 43.4°~17.3°
Aperture Type
2.8-12mm: Fixed aperture
8-32mm: DC drive
ICR Blue Glass support
fill light
IR Distance
Hybrid Fill Light (-X):
2.8-12mm: 40 meters (5 meters for human face)
8-32mm: 100 meters (face 14 meters)
White light (-L):
2.8-12mm: 40 meters (5 meters for human face)
8-32mm: 100 meters (face 14 meters)
Wavelength range mixed fill light (-X): 750nm
Smart function
Smart Events
Border crossing detection, area intrusion detection, entry/exit area detection, item left/taken detection, loitering detection, people gathering detection, fast motion detection, parking detection
Scene change detection, audio abnormality, audio steep rise/drop detection, audio presence or absence detection, virtual focus detection
Cross count support
Road Monitoring
Vehicle detection: support license plate recognition and capture, model / car logo / body color / license plate color recognition
Mixed-traffic detection: Detect vehicles driving in the forward or reverse direction, pedestrians and non-motor vehicles, automatically identify vehicle license plates, and capture pictures of vehicles without license plates.
professional intelligence (deep learning algorithm)
Facial Recognition
Number of supported face databases: 3
The number of faces supported by each face database: 30,000, and a single face does not exceed 300Kb
Smart City Management supports a 1-24 hour detection cycle, and supports the detection of illegal stalls
Vehicle Mode Body Color, Brand
professional intelligence (ordinary algorithm)
Alarm linkage motion detection, blocking alarm, network cable disconnection, IP address conflict, illegal login, memory full, memory error
certified
Protection IP68
optical port
Interface type throwing line
professional intelligence (face)
Face capture
Face detection performance: can support up to 30 faces at the same time, support detection, tracking, scoring
Support uploading large background images and small face images