Learning to Automatically Catch Potholes in Worldwide Road Scene Images
J. Javier Yebes, David Montero, Ignacio Arriola
Abstract
Among several road hazards that are present in any paved way in the world, potholes are one of the most annoying and involving higher maintenance costs. There is an increasing interest on the automated detection of these hazards enabled by technological and research progress. Our work tackled the challenge of pothole detection from images of real world road scenes. The main novelty resides on the application of latest progress in Artificial Intelligence to learn the visual appearance of potholes. We built a large dataset of images with pothole annotations. They contained road scenes from different cities in the world, taken with different cameras, vehicles and viewpoints under varied environmental conditions. Then, we fine-tuned four different object detection models based on Deep Neural Networks. We achieved mean average precision above 75% and we used the pothole detector on the Nvidia DrivePX2 platform running at 5â6 frames per second. Moreover, it was deployed on a real vehicle driving at speeds below 60 km/h to notify the detected potholes to a given Internet of Things platform as part of AUTOPILOT H2020 project.