Govt to use AI system to identify road defects from phone videos

The Government of India is reportedly testing an AI system that can identify road defects such as potholes and cracks from videos recorded on smartphones. The system was first tested in select locations and is now being experimented nationwide.

In Short

• Govt is testing AI system to identify potholes, cracks on rural roads
• System uses videos recorded on mobile phones
• It is now being tested as part of a nationwide exercise

The Government of India is testing a new way to identify defects such as potholes and cracks on rural roads, with the help of AI. It is said that the government has rolled out a nationwide exercise to examine this new system

As per a report from Hindustan Times, the government is testing a new AI system that can identify potholes, cracks and other visible defects on rural roads from videos recorded on smartphones.

The report adds that the Ministry of Rural Development has given “in-principle approval” for trials of the AI-based road-condition assessment system on roads built under the Pradhan Mantri Gram Sadak Yojana (PMGSY). The system is reportedly being used as part of a nationwide field exercise that began on September 1, after states and Union territories identified Programme Implementation Units for trials on PMGSY roads.

This AI system is being developed by the Centre for Development of Advanced Computing (C-DAC) with the National Rural Infrastructure Development Agency (NRIDA).

How will this work?

Currently, road maintenance under the scheme is monitored through physical inspections and geo-tagged photographs uploaded by field officials on the eMARG mobile application. Under the new trials, the system is being tested to identify seven types of visible defects – potholes, longitudinal cracks, transverse cracks, edge breaks, surface depressions, patches and vegetation-related obstructions.

The report states that road-condition videos are currently being recorded through an app using mobile devices already mounted on vehicles available with the implementation units. Dedicated dashcams may be considered later.

It is said that the AI-generated assessments can help identify and document maintenance needs. The technology is being tested only as a support tool for road maintenance decisions, and engineers will continue to physically verify defects and compare their measurements with the findings generated by the AI system.

Engineers to review AI assessment

If an AI assessment differs from an engineer’s assessment, the difference will reportedly be examined during the field-validation exercise, and a verification and review protocol will be finalised before wider implementation. The ministry is also examining whether roads can eventually be assessed every quarter under a proposed technology-enabled maintenance framework, although the frequency has not yet been finalised.

Existing contractual provisions, including those related to maintenance and the Defects Liability Period, are said to continue to apply. The Defects Liability Period is five years from the date of completion, during which the contractor remains responsible for correcting defects and carrying out routine maintenance.

The initial trials of this system reportedly began earlier this year, covering six roads in Pune, Lucknow district, Kamrup district in Assam and Ri-Bhoi district in Meghalaya. Later, an improved version of the model was tested on seven roads in Kanpur district and Berasia block of Bhopal district. The wider deployment would depend on the outcome of the national validation exercise, operational readiness and further approvals.

This reflects a broader use of AI in road management in India. Recently, the Delhi government has also proposed an AI-based traffic and safety audit of more than 420 km of roads in South Delhi to study congestion hotspots, bottlenecks and accident-prone stretches.

- Ends