Abstract
Potholes are a common nuisance that most people have had the displeasure of coming across. These bowl-shaped cavities in the road cause a large proportion of automobile related accidents, either directly or indirectly. Begetting the process of getting a pothole covered/fixed is a time consuming one that involves informing the appropriate authorities and having them take action. Implementing a system that involves citizens in the process of detecting pothole is what is being envisioned. The system includes a mobile application which is capable of taking a photo, this photo is then sent to a backend server where it is processed by a neural network model capable of detecting potholes. If any potholes are detected in the photo then the photo is saved along with its geolocation and forwarded to the appropriate authorities thus providing an efficient method of identifying potholes and removing the communication gap between the government and its citizens.
Highlights
Potholes are depressions which exist on the surface of a pavement that usually consist of a bowl-shape
A mobile application is developed that allows citizens to capture photos of potholes and send to a backend server which is equipped with a neural network model capable of detecting potholes present in the image
Every cell predicts a number of bounding boxes (B) along with a score of confidence which constitutes the intersection over union (IOU) and the ground truth bounding box
Summary
Potholes are depressions which exist on the surface of a pavement that usually consist of a bowl-shape. A mobile application is developed that allows citizens to capture photos of potholes and send to a backend server which is equipped with a neural network model capable of detecting potholes present in the image. The model is simple to construct and can be directly trained on full images This neural network model has been trained with 1500 images of potholes and is capable of detecting potholes from an image. [2] Uses a blackbox camera to collect data in the form of images and detects potholes using the proprietary algorithm and convey the information to road authorities in the city of Seoul, Korea. [5] Uses vibration meter and accelerometer along with gps module to detect potholes and inform to the authorities so that proper road conditions can be restored.
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call