An Intelligent Approach for Speed Bump Detection

In this paper, we propose an intelligent approach that uses YOLO algorithms to detect marked and unmarked bumps. The proposed approach for speed bumps based on real data collected from the Upper Egypt region in Egypt. Several data augmentations were performed on it. The proposed approach reached a mAP of 81% in the model while also calculating the distance from the car to the bump. We have used the Jetson Nano and the Xbox Kinect camera hardware to achieve this objective. Based on road preview using a camera, this work can be used to inform and alert the driver via a mobile application to make the ride more safe, prevent accidents, and reduce vehicle damage.