Despite significant enhancements in car safety technology, traffic accidents still happen, so finding creative ways to reduce the risk is necessary, particularly in situations with poor vision, like fog. This study presents a unique method using deep learning, namely the Faster MobileNet model, to reduce traffic accidents in situations where eyesight is impaired. The suggested solution makes use of a live video feed from the car’s front-mounted camera and Faster MobileNet to identify and highlight things on the driver’s display instantly. To improve image quality, the system incorporates image processing methods such as data augmentation, normalization, and scaling. Convolutional and pooling layers are used in feature extraction, together with a Region Proposal Network (RPN) to find potential object areas. Classifiers are used in object detection to recognize different kinds of objects and give the driver pertinent information. Non-Maximum Suppression (NMS) is incorporated into the implementation as a post-processing step to eliminate duplicate detections and improve the result. By acting as an intelligent helper, the created application greatly increases poor vision drivers’ visibility and promotes a safer driving environment.
Enhancing Road Safety in Diminished Vision Conditions: A Deep Learning Approach Using Faster MobileNet for Real-time Object Detection
2024-06-21
1077992 byte
Conference paper
Electronic Resource
English
Enhancing Road Safety: Road Friction Object Detection
IEEE | 2024
|Enhancing Road Infrastructure Maintenance Using Deep Learning Approach
Springer Verlag | 2024
|Forecasting Fire Using MobileNet Architecture
IEEE | 2023
|