Static scene target detection based on VIBE algorithm
Mingrui Yang, Qunyi Chu · 2021 2nd International Seminar on Artificial Intelligence, Networking and Information Technology (AINIT) · 2021
In this paper, we analyze the foreground target extraction problem of surveillance videos and establish various mathematical models for different types of surveillance videos: two-frame difference model, Single Gaussian Model (SGM) and VIBE model, and apply these models comprehensively to achieve the extraction of foreground targets in surveillance videos with different background characteristics. On the basis of foreground target extraction, the effective features of the image sequence are extracted to determine the abnormal behavior of the crowd in the video.