Improved Obstacle Detection and Alert System for the Visually Impaired Using Image Processing and Bluetooth Audio to enhance safe and independent mobility

Aim: The present research focuses on developing an Improved Obstacle Detection and Alert System, combining techniques of image processing with Bluetooth audio technology for enhancing visually impaired people's safety and independence in dynamic environments. Materials and Methods: The system is categorized into two groups. Group 1: Existing System employs a camera-based system which uses YOLOv3- Tiny model to detect possible dangers of surrounding objects. Group 2 refers to the MobilNet SSDv3 model which detects the surrounding obstacles with a Bluetooth audio alert system. Result: The system exhibited an increase in accuracy relative to the YOLOv3-tiny model. The proposed system MobileNet SSDv3 model mean is 94% while the existing system YOLOv3-Tiny model mean is 90%, Which is lower than the proposed model. The independent samples test verified that this gain increase was statistically significant at 0.036. Conclusion: This work shows that the obstacle detection system using MobiiNet SSDv3 provides reliable detection and improved performance, enabling timely audio alerts for the visually impaired.