Huitong face recognition integrated server platform

  The security monitoring system in the comprehensive promotion of technology, the front camera transmission mode and format, and the backend server computing storage capacity replacement, there has been a qualitative leap, from the original to see what happened, to see what is now done who at what time, how to do. Security monitoring system has made great progress, however, it is still only at the technical level to upgrade, and did not really realize intelligent, as the limbs are strong but easy to move the robot. In the control room backstage, still need administrator staring at the monitor screen, for fear of missing any alarm information; the process of handling cases, the police still need a lot of cost check over the video record, for fear of missing any important clues.

  In order to effectively solve such a heavy and inefficient labor, improve the effective use of the police force, reduce the burden on the police. We need a set of intelligent security monitoring system to replace heavy single repeated labor. Huitong technology face recognition system not only to the security monitoring system of intelligent resolution and recording, providing conditions retrieval function, make search query more smart and simple; greatly improve the operational efficiency of the security monitoring system, reduce the burden of the investigators, shorten the time of investigation, to improve the efficiency of case.

Face recognition intelligent front end

  Face recognition intelligent front-end is composed of a protective cover and intelligent HD cameras, intelligent front-end shield protection level of IP66 with dust, waterproof, anti surge drop function; built-in camera with 2 million pixel high-definition progressive scan CMOS, high definition, low intensity, high frame rate and high color reproduction characteristics. All products have independent intellectual property rights.

Main characteristics

● Uses a 1/1.8 inch 2 million 300 thousand pixel high-definition camera intelligent progressive scan CMOS;

● Also supports H.265 and H.264 coding, the maximum output of HD 1920*1080@25fps Full real-time image, ultra low delay, ultra low bit rate;

● Using ROI, SVC and other video compression technology, high compression ratio, suitable for flexible processing, ultra low delay, ultra low bit rate;

● Supports dual stream output at the same time, can output H.265 and H.264 stream dual stream can be independently set up different HD resolution, frame rate and bit rate;

● Supports intelligent program curing, support maximum 64G TF support local storage, capture images broken network http;

● Supports 2D/3D digital noise reduction function;

● Face detection, support l classification of non motorized vehicles (age, gender and glasses) and capture; support personnel statistics technology;

Product advantage

1) The use of Huitong company in the camera hardware advantage, front-end development of specialized intelligent face recognition, and the introduction of the face region in the image quality of camera exposure, significantly enhance the accuracy of face recognition. Internal tests show that the accuracy rate can be increased by 5-10% compared with the normal camera.

2) Compared to the ordinary camera, face recognition intelligent front end of the output image without H.264 codec compression loss, image quality is higher, the normal face recognition accuracy can be increased 3%~5% (internal test).

3) Rich experience in construction and installation of construction materials and guide. Through the professional construction, installation, debugging, to ensure the quality of human face imaging, to improve the accuracy of human face.

4) Has the independent patent facial feature recognition and tracking algorithm, the front-end equipment with intelligent, makes the quality to capture images is more suitable for face recognition system.

Typical applications

Technical parameter

Intelligent server

  The intelligent server is composed of hardware platform and integrated management service platform. Intelligent server is mainly used to have independent intellectual property rights of the face modeling algorithm, the face matching algorithm, the software architect of the two development of the integrated management service platform. Powerful hardware platform to support intelligent agile service platform, and the core algorithm of high efficiency and the ability to play the hardware platform.

Technical parameter

Face real-time alarm server

  The server realizes face capture images in HT-AS1500-FA face real-time alarm (by intelligent front-end, time, age, gender, wear glasses and other attributes) storage and query, real-time blacklist, alarm and warning function personnel wandering. The HT-AS1500-FA server receives the real-time alarm face face image intelligent front end capture through the VPN special network, through the comparison with the pre entry blacklist database matching, similarity and draw the corresponding picture, and the similarity beyond the cordon picture identification and alarm.

Main characteristics

● Supports up to 16 access road face capture machine, access blacklist real-time comparison.

● The number of L blacklist recommended less than 1000 people.

● Support for access face capture machine to capture images of gender, age, whether wearing glasses and other attributes of the query.

● Supports face capture image query, modeling, matching identification and retrieval functions.

● Humanized operation, simple and convenient, to support the black list of bulk import.

● Support images, video, real-time images and other ways to register a black list.

● Supports human face repeat function.

Product function

  Face inquiry function: capture images can be realized on the face snapshot camera capture images query. You can choose single or multiple capture machine capture photos query; you can query the different time at the same time, snapping photos; support gender age, whether wearing glasses human attributes capture photos query.

  Blacklist real-time comparison alarm function: the system may need to be dispatched personnel information (including name, gender, ID number, home address, face photos and other information) added to the blacklist database. Then according to the time, location, grade, degree of control met the alarm threshold and other information, protection of personnel. The system of opening position in certain key places (such as ports, docks, hotels, airport entrance) to capture real-time and blacklist database who dispatched officers to conduct real-time comparison, if the face similarity reaches the set alarm threshold, automatic warning system through sound and alert monitoring management. According to the monitoring and management personnel can double-click the alarm information view video verification and image capture.

Face repeat function

  That system will face photos for each snapshot were recorded when the repeated number reaches a certain threshold, it will face repeated alarm.

Face capture retrieval server

  HT-AS1500-FRS face capture retrieval server can realize face capture front-end and with face detection function of intelligent IPC camera access, solve intelligent face recognition front-end access number effectively the problem of excessive, can realize the face image database queries to capture, capture library face photo retrieval and capture library check (single repetition and support the double function of repetition). Can be widely used in the criminal investigation and traffic industry, through face recognition intelligent front end capture photo retrieval query the various surveillance area, fast locking track suspects.

Main characteristics

● Provides back-end intelligent video analysis focused on the solution.

● Supports up to 64 access road face capture machine or face detection function with Smart IPC.

● Support for access face capture machine to capture images of gender, age, whether wearing glasses and other attributes of the query.

● Support for the library to capture the face image retrieval, support instant retrieval and retrieval tasks two retrieval mode

● Support library to capture single and double repetition repetition.

● Supports 1 pairs of 1 pairs and 1 pairs of multiple alignments.

Product function

  Capture library search function: in the massive personnel information, enter a search face photos, to compare other face image filter and portrait feature point alignment algorithm and the capture of the library, according to the similarity in order to help the police can quickly find accord with the conditions of personnel information retrieval.

  Capture repetition functions: divided into single and double repetition repetition. Single repetition refers to the repeated search personnel in the same capture library; double repetition refers to two different capture library, search personnel appear at the same time in two in the library.