Pothole Detection using CNN and YOLO v7 Algorithm
E Sai Tarun Kumar Reddy, V Rajaram
Abstract
Various factors, which include excessive speed, driver distraction, failure to wear safety equipment, and road hazards such as potholes, account for the majority of fatal and injury-causing incidents in India. The number of potholes is likely to be responsible for most of these accidents. Vehicles and people are both negatively impacted by potholes, which are the imperfections on roads. Whenever there is an increase in the amount of traffic volume, so does the number of pothole incidents. India has been traditionally relying heavily on automobile travel. According to a recent survey, there have been over 295 million registrations of vehicles to date, and that number continues to rise rapidly. Several methods for detecting potholes have already been developed, which include the use of sensors and the attachment of specialized hardware to automobiles; however, these methods are either too expensive or too difficult to be widely adopted. As a result, a new method has been devised in this proposed work that utilizes a smartphone's camera and current location to categorize the potholes and determine the pothole count. The YOLO v7 algorithm is utilized for pothole categorization, and the Google API and Accelerometer are utilized for pothole counting.