• Deep learning algorithm is adopted, based on massive pictures and video resources, and target features are extracted by the machine itself to form deep learnable human clothing images for recognition
• Based on pedestrian trajectory analysis, the entry and departure of target personnel in the specified scene are counted, and the passenger flow is counted
• On the basis of passenger flow, it also supports clothing mode or face mode switch: clothing mode: 1. It supports comparison and weight removal based on human body and attributes, and filters the passenger flow count of custom clothing and attributes. 2. Support the classification of human attributes, including: gender, age, carrying things, hats (helmets), masks
• Face mode: Support face capture, Angle filtering, face attribute classification, support dynamic deweighting based on face comparison
• Event alarm function: Support motion detection, block alarm, hard disk full, hard disk error, network disconnection, IP address conflict, illegal access, abnormal restart
• Support the standard 256 GB MicroSD/MicroSDHC/MicroSDXC card storage, support 10 M / 100 M / 1000 M adaptive front-end ports, support wi-fi (- W type support)
• Support open network video interface, ISAPI, GB/T28181-2016, ISUP5.0, video library
• Supports three-level user rights management, authorized users and passwords, and IP address filtering