Chongqing high-speed video surveillance can realize automatic identification

At present, it is mainly to observe the video surveillance image manually. Some illegal accidents have to be replayed from the monitoring in order to obtain evidence, which is troublesome. Recently, a digital image automatic recognition and processing system for highways, plazas and other violations passed the acceptance test and was applied to Chongqing's major expressway sections. Without manual monitoring, you can automatically pop up and save when you encounter a violation screen, so you can find it at any time.

Existing highway video surveillance video is mainly monitored by humans, of which more than 90% of video surveillance images cannot be focused. Failure to detect operational security incidents at the first time is a major bottleneck in existing video surveillance systems, and event discovery coverage is extremely low.

This digital image automatic recognition and processing system was developed by the Chongqing Municipal Institute of Transportation IoT Engineering Technology Research. A set of faulty vehicles, illegal parking, and pedestrians facing highway tunnels, key road sections, and toll plazas were established according to the characteristics of various scenarios. Automatic detection methods and early-warning technologies for spills, fires, etc.

For example, the front camera photographed a group of illegally parked vehicles on the highway. According to the past, it must be monitored at the backstage of this time, and just to see this scene, in order to be aware of this situation in a timely manner, otherwise it can only be understood afterwards. However, this new system does not require manual monitoring in the background. As long as the camera captures the screen of illegal or faulty vehicles, the intelligent analysis and processing system in the background can automatically identify and form image pop-ups and automatically save them. At the same time, an early warning signal was sent to the relevant departments to handle the incident.

At present, this scientific and technological achievement has been applied to a number of typical sections of the Chongqing Expressway, which enables automatic screening and filtering of surveillance images, increasing the coverage of operational security incident detection from the original 10% to more than 90%, and covering incidents within the scope of events. The detection rate is over 80%, and the monitoring range deviation detection rate has reached over 85%, which ensures the reliability of the monitoring system and will play an increasingly important role in ensuring the safe operation of the city's traffic.