Children’s Accident Prevention System Based on Real-time Video Interpretation and GPS Information

This paper proposes a smart badge system for children's safety. We implement semantic segmentation technology, GPS-based location tracking, and safety zone setting functions. First, we implemented a situational notification function using semantic segmentation technology. This system analyzes the child's surrounding environment in real-time to detect dangerous situations and provides immediate notification through a badge for children. This can raise children's traffic safety awareness and prevent jaywalking. Additionally, semantic segmentation technology, which can accurately recognize and classify objects in images, dramatically improves the accuracy of situational judgment. This can reduce unnecessary notifications and provide reliable notifications about actual risk situations. GPS-based real-time location tracking and safety zone setting functions were implemented. The child's real-time location can be checked on the guardian app through the GPS sensor built into the child's badge. This allows continuous monitoring of the child's safety status. Additionally, setting a child's safety zone in the guardian app immediately sends a notification to the guardian when the child leaves the safety zone, enabling crime prevention and rapid response. GPS-based technology improves the user experience by making it easier for guardians to determine their child's location. This system can be applied to various environments and situations and expanded to other child safety services. Additionally, the safety zone setting function helps relieve guardians' anxiety. This can be said to present an innovative solution to child safety.