Towards Vision Based Bi-Cycles Detection and An Autonomous Decision Making System for Driving Situation Awareness

In response to the growing global emphasis on healthy living and environmental consciousness, the popularity of eco-friendly transportation options like bicycles has surged. However, cyclist safety remains a critical concern, particularly in interactions with automobiles, where accidents often result in severe consequences. This project aims to enhance cyclist safety through machine learning-based bicycle detection and situational awareness techniques, aiming to mitigate vehicle-bicycle collisions. By integrating OpenCV lane detection and YOLOv8-based vehicle/bicycle detection with distance estimation, the system can analyze real-time road conditions and make informed decisions, including speed reduction, overtaking safety assessment, and maintaining safe distances from cyclists. Testing validates the system's accurate detection, precise analysis, and effective decision-making capabilities, ensuring vehicles respond appropriately to encountered road scenarios. Ultimately, this project addresses the pressing need for cyclist safety in vehicle-rich environments, leveraging machine learning, real-time analysis, and advanced detection technologies to foster safer roads.