Any Questions?: +86-23-67305242
Chat Now



Magic Cube Intelligent Analysis Box Product Specification

Product Features 
Device Access. 
◼ Access to ONVIF, RTSP-compliant HD network cameras
◼ Access to face capture machines that comply with GA/T1400 protocol
◼ Supports H.265/H.264 video standard and up to 8 megapixel IPC for access intelligence applications. 
◼ Total library supports up to 300,000 face images, supporting 64 libraries
◼ Multi-class algorithm mixing: face mode, mixed mode (face-human binding + video structuring)
◼ Face recognition: 16 video streams supported
◼ Video structuring: 16 video streams supported
◼ Smart Alert: Supports 16 video streams
◼ Support face capture, face recognition comparison alarm, stranger recognition alarm, etc.
◼ Support recognition of face attributes: gender, age, wearing a hat, glasses, mask, etc.
◼ Support face, human body, motor vehicle, non-motor vehicle, license plate capture and face-human binding, human-non-motor vehicle binding
◼ Support human body attributes such as upper and lower clothing color, upper and lower clothing style, backpack status, and whether to wear a helmet analysis
◼ Support motor vehicle classification, color, brand, driving direction and non-motor vehicle classification and other vehicle attributes analysis
◼ Support license plate recognition and license plate - motor vehicle binding relationship
◼ Support people and vehicles separation, mixing line crossing, area invasion, vehicle illegal parking, over the line counting and other intelligent alert function system features. 
◼ Support MegConnect V2.0 protocol to access Cloudwing, Cloudbridge, Pangu, and other software platforms
◼ Available via GA/T1400 standard protocol cloud or third-party view library  platforms
◼ Provide a rich RESTful API protocol interface for thirdparty platforms to do docking development application scenarios. 
◼ Provide the ability of face recognition and full target structured analysis, with the upper platform software can realize the closed loop of face capture recognition and video structured scenes.
◼ Applied to places that need face recognition, such as entrances and exits of intelligent communities, entrances and exits of office buildings, key personnel list control and identification alarm, stranger identification, etc.
◼ It is applied to structured places that need face man-machine non-plate capture and license plate recognition, such as people and vehicle control in smart parks and license plate recognition in smart communities.
System Parameters
Main Processor High-performance embedded microprocessors
Operating System Embedded Linux
Device Access
Video Stream Input Video resolution. 1920 x 1080 (2 million), 2560 x 1440 (4 million), 3840 x 2160 (8 million)
Video decoding type H.264/H.265
Face Capture Machine Maximum support for 32-way capture machine access (capture machine should support GA/T1400 protocol)
Intelligent Parameters
Working mode: Mixed mode (Algorithms in parallel, configurable by channel, flexible switching between face, structured, and vigilance algorithms) Face person binding + face recognition: (maximum 16 ways) Face capture, face recognition, face attributes, human body capture, human attributes, face human binding
Video structuring: (up to 16 channels) 1) Image capture: face, human body, motor vehicle, non-motor vehicle, license plate 2) Attribute output: face, human body, motor vehicle, non-motor vehicle, license plate 3) License plate recognition 4) Relationships: face-person binding, car-plate binding, personnon-motor vehicle binding;Smart Alert: (max. 16 channels) 1) Separation of people, cars and non-motorized vehicles, highdefinition capture 2) Output of alarm events such as mixing line crossing, area intrusion, vehicle parking violation, and line crossing count
Proactive reporting 1) Face capture, face recognition, face attribute results 2) The face man-machine non-plate capture and attributes, license plate recognition results 3) Alert Alerts
Human face Capture rate ≥ 99%
False capture rate < 1%
Whitelist recognition pass rate > 99.5%
Recognition misidentification rate: < 0.5%
The largest library: 300,000 portrait library
Human body Capture rate ≥ 95%
False capture rate < 1%
Motor Vehicles Capture rate ≥ 90%
False capture rate < 1%
Non-motorized vehicles Capture rate ≥ 95%
False capture rate < 1%
License plate Capture rate ≥ 95%
False capture rate < 1%
Recognition rate ≥ 95%
Interface parameters
Network Interface 2 pcs, 100M/1000M Adaptive Ethernet, RJ45 interface
Alarm input interface 4 way switching
Alarm output interface 4 way switching
Audio Output 1 Road
Audio Input 1 Road
Front USB port Front 1 USB2.0, 1 USB3.0
Rear USB port 2 rear USB2.0
RS485 2 Roads
Reset button 1 pc
Power indicator (PWR) 1 pc
Run indicator (RUN) 1pc, two-color (red + green)
System Functions
Face Recognition Entrance and exit staff passage; key personnel list control and identification alarm, stranger identification, etc.
Video structuring Face, human body, motor vehicle, non-motor vehicle, license plate capture, attribute analysis, license plate recognition
Dual network port Support "multi-access setting", "load balancing" and "master/standby mode"
Log query Enables query, search and display of capture information of face man-machine non-tag
Environmental requirements
Operating temperature -20℃ ~ +60℃
Storage temperature -30℃ ~ +70℃
Relative Humidity 10% ~ 90%RH, non-condensing
Power supply DC12V±10%, 2A
Structure Metal Case
Dimensions (length, depth, height) 229×193×49(mm)
weight <4KG