Pothole Detection Using a Self Driving Car

With the growing popularity of self-driving cars on roads, the assurance of security and functionality is an urgent need. In that respect, detecting potholes is an important piece of research, because they are hazards not only for vehicles being driven autonomously but also those being steered by humans. This paper proposes a novel way for detecting potholes with its hardware setup including an Arduino and a Raspberry Pi, where the Arduino acts as a Slave device and Raspberry Pi acts as a Master device. Furthermore, we use the TinyYOLOv4 which is a state-of-the-art object detection model for training a large scale pothole detection dataset..