BI-CN/18-20-0001: Research on Intelligent Early Warning of Emergency from Perspective of Public Security

Outline of national development plan for science and technology in medium- and long-term points out that public security is the cornerstone of national security and social stability. We are facing big public security threat in China, which proposes important strategic requirements for science and technology. To cope with this challenge, our government needs to construct one technology system of public security to prevent or control the emergency and adopt efficient response effectively. To bring about a multi-functions integrated emergency guarantee pattern with information and intelligence technology application as the leader, this system will help our government to strengthen its capability of dealing emergencies with rapid response and by predicting sudden events scientifically. From perspective of public safety, this project utilizes video data captured by cameras installed in public areas and does early warning for crowd abnormal activities, as well as providing service for public safety of ethnic minority areas.

How to get new situations or new problems accurately immediately based on information technology and adopt active and effective measure to reduce losses will contribute to harmonious social development, promote development of national unity, public security and social stability. If we get emergency information supported by computer vision in the first place based on videos captured by many cameras installed in public places, and issue warnings immediately, we will reduce the loss of life and property of the public with immediate protection measures. If we cannot finish real-time analysis especially occurrence of danger or abnormal events and cannot detect actively and do some early warning, then it would be a waste of video resources. To maintain the unity and stability of our public security and society, our team by two sides tries to utilize videos captured by cameras installed in public places fully and recognize abnormal crowd behaviors based on computer cognitive computing to do early warning for emergencies for public security with requirements of intelligent video surveillance and advanced I/O multimedia equipment.

Although there are many important research theories and applications in video behavior recognition in recent years. But there are only a limited number of research works related abnormal behavior based emergency from perspective of public security, especially for the performance of real-time or robustness of abnormal crowd behaviors recognition algorithms. There is still a lot of research work to do in early warning for sudden group abnormal behavior. Therefore, research on intelligent early warning of emergency from perspective of public security has a certain frontier technical difficulty and exploration with important scientific significance and research value.