The current process to maintain road assets is a time-consuming, manual labor-intensive process, error-prone and costly. Computer Vision will eliminate the limitations of the traditional survey process & deliver efficiency across the value chain. This will result in better response time to fix road asset issues, making roads and highways safer.

Section 1: A revolution in the process of maintaining roads

Right now, roads are being surveyed manually and the data is entered into a computer after the fieldwork has been completed. This requires several manual workarounds to be performed in case of any shortcomings of the process, like, for example, loss of data, system failure, etc.

After the field data has been collected, it is immediately entered into the computer for reporting purposes. For example, the average response time to report road defects from the last 3 months was about a month!

The software can help us retrieve the data from the system and present it to us quickly. However, this can also be time-consuming and error-prone.

Computer Vision removes many of these constraints and produces a much better-organized, more accurate and shorter report.

How Computer Vision will make it easier

Currently, road survey is carried out manually, with surveyors physically riding along the highway. The surveyor would manually count vehicle numbers on the highway to confirm vehicles traveling in the traffic lanes.

Computer Vision can provide answers with just a few mouse clicks, eliminating the need for a physical survey.

Many attributes can be computed with computer vision, such as perimeter driving pattern/traffic flow/vehicle count/grid layout/vehicle direction/need to drive, etc.

Many other problems can be addressed, such as identifying expired number plates or lost and found documents.

Computer Vision can also provide answers about, "Where are all the vans/trucks traveling during peak hour? I need to monitor a specific road segment with maximum trucks & vans per hour.

What is Computer Vision?

Computer vision is the ability of a camera to analyze an image to understand the various elements of the image. This enables computers to identify objects, measure distances and create 3D models of objects in a virtual environment. Computer vision is similar to the concept of the human eye or our brain. Computer vision relies on image recognition, which is similar to how the human eye interprets visual information. This concept is similar to the work of Amazon's Amazon Web Services (AWS). When you use Amazon's Echo device to search on Amazon, you are interacting with a computer vision and machine learning, which tries to understand the images you are looking at. The advantage of this technology over other technologies is that it will be faster, simpler and cheaper.

A future with safer roads

Tech leaders are working hard to develop new Artificial Intelligence (AI) platforms that will help analyze the content within the image, identify signs, obstacles, driving conditions and other opportunities for resolving operational issues. It also helps the transport industry to identify the overall condition of roads for predictive maintenance and lower maintenance costs.

In India too, a consortium of start-ups and the government has developed the first such platform called Mobile Open Data Set (MODS) for delivering on an effort to make the best use of Open Data and create a road map for leveraging this information for planning & monitoring of various road infrastructure projects across India.

Conclusion

• Road assets serve as our arteries to the economy.
• By building more roads, our economy will develop, requiring more cars and more vehicles.
• Computer vision could assist in identifying defects, deformations & road users.
• With more cars and more road users, our roads will become congested & less efficient.
• The study showed how computer vision technology could be used to detect issues, streamline asset inspections and improve response time.

For more insights, please refer to the following:

Intelligent Surveillance Solutions: How NIMC is pioneering machine learning for accurate data collection & preventive maintenance of traffic assets.