Project overview

Through the on-site video analysis application system, a large number of personnel will be required to visually observe and screen frame by frame. Through automated processing by the system, the risk of missed viewing and screening during manual investigation is solved, reducing physical harm to video surveillance police officers. At the same time, by analyzing the correlation between the portraits, vehicles, features, behaviors, and other information appearing in the video and case management, an evidence chain is formed. And perform image enhancement and video slicing techniques on the images in the video to identify key nodes and slice the images and data.

Analysis function

1. Main functions: on-site video management, vehicle detection, personnel detection, behavior analysis (trajectory tracking, gathering, banner pulling, sign lifting, kneeling, fighting), image enhancement, video slicing, and identity recognition.

2. A complex AI application system consisting of over 20 parallel and interleaved AI modules, analyzing video sources from 100000+cameras. Support video streams captured by multiple devices; Support two analyses on the analysis results; Support cross camera tracking.

3.The efficiency of video analysis and search has increased by 20 times.

      4.Real time monitoring efficiency has increased by 20 times, with one police officer monitoring 140 cameras in real-time (previously only 7 at most).

Greatly reduced the work pressure on police officers.