Detection of Potholes Using CNN

EasyChair Preprint no. 11043

Abstract

Without a doubt, roads are the ones that connect people from one place to another without any hiccups. India plays a vital role in economic maximization as it serves as the second-largest road network. We have a delay problem that causes road accidents like Tyre bursts and collisions due to improper maintenance, resulting in the occurrence of large cavities as the road becomes worse, even more in rainy conditions, forming a lot of potholes on it. Deep Learning is a technique that learns features and tasks directly from raw data.To create a deep learning model that will aid in pothole detection . Uses a video as an input to build a CNN model that uses OpenCV to spot potholes in real-time. This project can be used to locate potholes and evaluate the condition of the road by both motorists and government representatives. In order to enable real-time vehicle operation (for driver assistance or autonomous driving) or offline data gathering for road repair, methods for spotting potholes on road surfaces are being developed. These factors have led to extensive research into pothole location techniques in studies carried out all around the world.

Keyphrases: CNN (Convolutional Neural Network), deep learning, Potholes

@Booklet{EasyChair:11043,
  author = {P Nithiyasree and T Arundhathi and S Soorya and Vatturi Harshitha and Kavitha Subramani},
  title = {Detection of Potholes Using CNN},
  howpublished = {EasyChair Preprint no. 11043},

  year = {EasyChair, 2023}}