Language: ChineselineEnglish

News

How to use urban monitoring? No need to

The face recognition camera can't see the face, what should we do? Face recognition has surpassed human beings in accuracy since 2015, and the recognition accuracy has jumped 100000 times in four years. In the future, more complex application scenarios will be unlocked, and the efficiency will be improved by several orders of magnitude. However, in the current practical application scenario, a single camera can not cover all areas, and the detection range of multiple cameras rarely overlaps. Not only that, limited by the installation height and density of the video monitoring probe, more photos are taken of the side face, the head or the back of the head of the pedestrian.
Before the realization of face recognition, we are faced with a more difficult problem: cameras are all over the city, and few cameras meet the face acquisition standard except for the cameras used by bayonet cameras and electronic police. The existing front-end equipment is difficult to meet the requirements of the height and angle required for face recognition, and needs to be replaced or transformed, which involves a long period High input construction. In order to tap the potential of the cameras that have been set up in the safe city and Xueliang project, other technologies need to be added.
AI + security industry leaders began to focus on the application of Reid (pedestrian recognition) technology and face recognition. Reid technology does not have high requirements for environment and equipment. It can make full use of existing equipment and infrastructure, and as long as certain pedestrian characteristics are identified, it can realize continuous tracking and track restoration of retrieval objects across cameras. Once face information is found in this process, the identity of personnel can be clear. This is a very ideal technology for public security.
For example, if a suspect's photo is known, Reid can collect data from the video database according to the photo to connect the suspect's tracks in each camera, which will provide great help to criminal investigation.
In general, for the field of monitoring, manufacturers hope to associate and find the same object in different videos through Reid technology. Only to achieve this degree, the whole monitoring operation can be called complete and see the overall situation.
It should be pointed out that Reid itself is also a very difficult technology. To identify the same person from different videos, video lighting conditions, monitoring area resolution and angle, target occlusion, and even people wearing similar clothes will cause difficulties in recognition. Data acquisition is very difficult, which will pose a great challenge to the algorithm. Now, the industry tries to work at the algorithm level to improve the recognition effect of Reid.
Old security companies such as Haikang and Dahua, AI four dragons, Huawei, and traditional Internet giant bat all have their own layout and accumulation in Reid field. Cloud has surpassed Alibaba, Tencent and Chinese Academy of Sciences in Reid technology, breaking three world records and pushing Reid recognition rate to a commercial level.
Conclusion: "brush face", as an important application in the field of computer vision, has swept the whole industry, and "know people" will promote the computer vision industry to enter a new stage of development. At present, Reid technology is mainly used in the field of security. The security industry has made in-depth research in many subdivisions of Reid, such as pedestrian detection, attitude estimation, behavior action recognition, pedestrian structured information extraction, etc. In the future, Reid and face recognition can be applied to more and more rich scenes, so that you can quickly experience the upgrading of intelligent technology in security, automatic driving, home life and other aspects.

CONTACT US

Contact: Micah.Zhang

Phone: 400 6400 262

Tel: 86-29-85535392

Email: info@okeyset.com

Add: No.4, Shixiang Road, Xi an City, Shaanxi Province, China

Scan the qr codeClose
the qr code