A major problem affecting cyclists globally, bicycle theft has arisen in an era of increased urban mobility and sophisticated technology. This paper presents a novel approach that combines real-time video analytics and machine learning algorithms for intelligent bicycle theft detection and surveillance. Our system monitors and analyzes suspicious activity around parked bicycles using a combination of image recognition and motion detection technologies. Which is a simple, low-cost bicycle theft control strategy that makes use of real-time vehicle theft and detection. In addition, the system has created algorithms such as an alarming algorithm and a short message trapping algorithm. The system detects potential theft attempts with high accuracy and low false positives by using YOLOv5s and DeepSORT for object detection and tracking, respectively, and CV2 for video and image capture. The video and image data are then processed in high-processing devices. Our solution also includes a dynamic alert mechanism that uploads the thief's image to the website in 1–2 seconds, notifying the owner of the bicycle and the local authorities in real-time. Extensive field trials are used to assess the effectiveness of the proposed system, showing a significant decrease in theft incidents and an enhancement in overall security. In addition to improving bicycle safety, this intelligent surveillance method offers a scalable model for urban security applications.
Bicycle Theft Detection by using YOLOv5s and DeepSORT algorithm
2024-12-07
793932 byte
Conference paper
Electronic Resource
English
Electric bicycle theft detection system and electric bicycle theft detection method
European Patent Office | 2024
|ELECTRIC BICYCLE THEFT DETECTION SYSTEM AND ELECTRIC BICYCLE THEFT DETECTION METHOD
European Patent Office | 2023
|ELECTRIC BICYCLE THEFT DETECTION SYSTEM AND ELECTRIC BICYCLE THEFT DETECTION METHOD
European Patent Office | 2023
|