Application of Face Recognition and Contrast Technology in Passage Gates

With the continuous advancement of technology, the role of video intelligent analysis in the field of security video surveillance has become more and more significant. The application of video intelligence analysis in rail transit mainly includes video quality diagnosis, video intelligent retrieval, face recognition and comparison, and intelligent behavior. Analysis and so on. This section focuses on the use of face recognition and contrast technology in gate turnstiles.

First, the system architecture

Big data face capture contrast system mainly uses face detection algorithm, face tracking algorithm, face quality scoring algorithm and face recognition algorithm. Realize the acquisition and modeling storage of the faces of import and export personnel of subway elevator entrances and exit gates, and the function of real-time blacklist comparison alarm and face retrieval.

Face capture machine

It is mainly deployed in the entrances and exits of subway elevators or gateway gates. Face capture machine built-in intelligent analysis algorithm to achieve the face of people entering and leaving the person to automatically capture, tracking, capture function. At the same time, it has automatic exposure function in the face area to capture the relatively clear face capture in the backlight environment.

Storage server, disk array

Store face snapshots, face photos, and various blacklist databases.

Stored and imported resident population face database, key monitoring personnel characteristics database, frontal personnel face feature database, stability-related personnel face feature database, national fugitives face feature database, * second-generation identity card photo library face Model storage, database redundancy backup.

The terminal can send a collected face picture to a management platform through a computer terminal, so as to confirm the identity information of the suspicious individual, and at the same time, feedback and compare the search results to the terminal devices at various levels.

Video analytics server

Contrast modeling: Modeling the face photo transmitted by the face snapshot machine. At the same time, the real-time and blacklist database personnel are compared.

Search service: Implements search comparison service for the person information database captured. To achieve the modeling of resident population database, key monitoring personnel database, forecourt personnel library, stability-related personnel library, national insolvency personnel library import, and updating to face storage server. Respond to the requests from the client terminals at each level for the face comparison search to achieve quick face search comparison.

Second, system application function

Blacklist real-time alarm, personnel dispatch function

The system can add information to the black list database as required (including name, gender, ID number, home address, face photo, etc.). Then, according to information such as time, location, dispatching level, and alarm threshold of acquaintance, personnel are armed. The system real-timely compares the personnel captured in the locations of specific key locations (such as entrances and exits of subway elevators or entrances and exits of gates) to real-time comparisons with people in the blacklist database. If the degree of face recognition reaches the set alarm threshold, The system can automatically perform warnings through sounds and other methods to alert monitoring and management personnel. The monitoring and management personnel can check the captured original image and video according to the double-click alarm information to verify.

Rapid identification of personnel

In the system, the face photo to be queried is input, and the system automatically detects the face information in the photo and intercepts the face. After the user selects the face to be retrieved, similarity, time period, and other parameters are set in the corresponding person information base. Within the search, the results of the similar faces retrieved at the end will be displayed on the interface, enabling rapid confirmation of the personnel information.

Personnel information collection and establishment of face information database

The camera captures, snaps, recognizes, and dispatches surveillance images to compare alarms. The face capture machine is capable of performing face detection and face tracking on the personnel passing through the set area, and the face quality scoring algorithm can automatically filter out a person's face image with the most clear front face information as the person's Take a picture. And upload face photo, snapshot location, capture time and other information to the face management platform for storage, unified storage, to facilitate the later search and query.