Local Chapter VIT Bhopal University
Coordinated byIndia ,
Status: Completed
Project Duration: 08 Apr 2023 - 05 Jun 2023
Open Source resources available from this project
Project background.
India is a country that is rising with an aggregate GDP of 7% every year, and as we are moving forward with a lot of advancement, there is a change in the road system in India. We have 6-lane highways, as well as Rural roads. However, despite these improvements, there are still many challenges that affect the traffic flow in India.
One of the most significant challenges is the impact of heavy rainfall on the road network. Indian cities often experience heavy rains during the monsoon season, which can cause flooding and waterlogging on roads, leading to traffic congestion and delays. This issue is compounded by the fact that many Indian cities have poor drainage systems, which can exacerbate the problem of flooding.
Another challenge is the high level of traffic congestion in many Indian cities. Even with the development of wider roads, the sheer volume of vehicles on the road can lead to bottlenecks and slow-moving traffic, particularly during peak hours.
The condition of the roads themselves contributes to the traffic problems. Poorly maintained roads with potholes and other damages can cause delays and are potentially a safety hazard.
We propose to develop a machine learning-based model for a road inspection system that will automate the detection of road abnormalities, defects, and damages, where we can observe all these irregularities that are present in the Indian Roads that is, what are the causes for pitfalls and sinks, what type of things are participating in road flooding, how less number of lanes in the city and town are affecting the traffic.
The problem.
India, one of the world’s fastest-growing economies, has a constantly expanding road network that connects cities, towns, and villages. Unfortunately, India’s present road inspection procedures are cumbersome, labor-intensive, and often ineffective. The roads in India are heavily used and are prone to frequent damage due to harsh weather conditions and heavy traffic, leading to decreased comfort for drivers and decreased economic efficiency. Moreover, manual road inspections are often limited in effectiveness, leading to delays in maintenance and repair activities.
To address this issue, we propose to develop a machine learning-based road inspection system that will automate the detection of road abnormalities, defects, and damages. The system will use HD cameras to capture live videos of various road defects and issues and then analyze the captured data using the Matlab machine learning toolbox to train and test the network, allowing the system to categorize road flaws and indicate locations that need maintenance or repair, and provide recommended actions for the municipality to fix/correct the road issues.
The proposed system will help solve the problem of time-consuming and labor-intensive road inspections, leading to more efficient maintenance and repair activities. This will benefit the local community directly by improving road quality and making them safer for vehicles, pedestrians, and bicycles. Additionally, the system will reduce the costs associated with road inspections, allowing for a more effective allocation of resources toward other needed areas. Ultimately, the proposed system will contribute to the sustainable development of India’s road infrastructure, making it more efficient and safer for all.
Project goals.
Project plan.
Week 1
Research previous work + Data Collection
Week 3
Data Pre-processing + Data Preparation
Week 4
Exploring pre-trained ML/DL models
Week 5
CNN training and validation + Transfer Learning
Week 6
System testing and accuracy reporting + Building Dashboard to visualize the output
Week 7
Deployment
Week 8
Generating final report
Learning outcomes.
Heavy traffic on roads causes their surfaces to deteriorate every day. This will have an effect on both, the comfort of the driver and the economy. Municipalities conduct routine inspections to maintain roads as efficiently as conceivable. Therefore, we aim to create an AI-based system that will offer suggested steps for the municipality to take in order to rectify or correct road issues. The method is broken down into three key tasks: gathering data, training and testing data, and creating and testing dashboards.
The study objective is to automatically detect any anomalies on the road by using machine learning to examine and analyze various sorts of road faults. For this, we will design, construct, and test an inspection system. A camera is built within the system to gather video streams from various roadways with and without problems.
The participants of this project will be able to enhance their Artificial intelligence and machine learning skills. They will gain a better understanding of the road system as well as the local municipalities and will also be able to understand the importance of team work.
