Developing a Pothole Detection Model for Pavement Surface Using Vehicle Big Data

Key Points

• The central aim is to develop a model for detecting potholes and pavement damage using vehicle big data.
• Developed a pothole detection model using data from accelerometers and vehicle sensors.
• Applied the model to municipal roads over an 11-week period.
• Evaluated detection accuracy for repair needs in pavement surfaces.
• The model successfully identified areas needing repairs with an accuracy of 40-60%.
• Confirmed that certain non-repair areas could be excluded after a specific period.
• Detected potential repair sites that may require future attention.