Warning Zone Abnormal Behavior Detection: Pre-Alarms for Risk of Exhibit Damage

In recent years, the vandalism of art exhibits has become a significant issue in museums and exhibitions. To address this problem, we propose an automated module system : Geo-Fence Discriminator (GFD), Behavior Surveillance (BS), and Behavior Response Alarm System (BRAS)—that predicts visitor behavior and ensures the safety of exhibits through preemptive warnings. The proposed system includes the GFD module, which determines safe and risky distances from the exhibits based on the coordinates of objects. For behavior recognition, the BS module collects a behavior dataset and classifies and identifies actions through trained image classification. Detected abnormal behavior information is then transmitted in real-time to BRAS, which displays alerts. This enables real-time prediction of abnormal visitor behavior near the exhibits and sends warnings to exhibition staff, thereby enhancing the protection of the artworks.