Community Blog Pothole Detection System: AI-based Assessment of Roads

Pothole Detection System: AI-based Assessment of Roads

This project is from the team ABESDevelopers, which was awarded third place (shared) in the Alibaba Global AI Innovation Challenge 2020.


Project Introduction

Pothole Detection System is a solution to automate the surveillance of roads (especially potholes) to help speed up the process of assessment of roads, which in turn would further lead to faster maintenance of roads

Target Problems

According to the data provided by the Ministry of Road Transport and Highways for India, over 9300 deaths and 25000 injuries were caused by potholes in the years 2015, 2016, and 2017 collectively. Additionally, potholes are responsible for a considerable amount of damage to vehicles. Patching potholes itself is a relatively straightforward process, but managing it around vehicular traffic can be complex, especially for heavily congested areas.

However, it is a time-consuming process to inspect and locate potholes, which leads to further delay in their patching. Our solution reduces this inspection time of roads for potholes and identifies the location of places where there are potholes. Hence, their patching can be done considerably faster.


We leverage the use of mobile sensors, such as accelerometers and gyroscopes, analyzing their readings and detecting patterns with the help of machine learning to detect any unusual behavior in them. Our model will then be able to use the sensor data in real-time from a mobile device to collect readings and classify whether a pothole is present there or not. Additionally, if a pothole is detected, then its location in the form of latitude and longitude is stored.

Alibaba Cloud Products Used

Technology Highlights

We used Object Storage Service (OSS) to store our data to train our machine learning models. Then, we used Machine Learning Platform for AI (PAI) to apply different classification machine learning algorithms on our data from OSS. These tools were easy to use and highly user-friendly.

About the Team

I am Aditya Gupta, the team leader for my team ABESDevelopers. My teammates include Harsh Agrawal and Aryan Sharma. We all are final year computer science undergraduate students from ABES Engineering College, Ghaziabad, India, which is affiliated to Dr. A.P.J. Abdul Kalam Technical University, Uttar Pradesh, India. Our passion to solve problems using technology has motivated us to participate in this competition.

0 0 0
Share on

Alibaba Cloud Project Hub

113 posts | 22 followers

You may also like


Alibaba Cloud Project Hub

113 posts | 22 followers

Related Products