How artificial intelligence, drones and cameras ensure the safety of roads and bridges

“It is a dangerous thing, Frodo, to go beyond the threshold: it is worth stepping on the road and, if you give free rein to your legs, it is not known where you will be taken .”

- DRR Tolkien, Lord of the Rings

The roads of Europe are the safest in the world. Current figures  show that there are 50 deaths per million inhabitants, compared with the global figure of 174 deaths per million. Despite this, every loss remains a tragedy. In 2017, 25,300 people died on European roads.

The causes of these accidents can be both human errors and weather conditions, and damage to the roadway and bridges. Although some things are beyond control, the condition of roads and bridges is a variable that can be controlled.

As soon as the road is paved, a combination of traffic and weather conditions begin to worsen and destroy the surface. Undetected cracks, potholes, or defects can quickly lead to more serious problems, such as costly car repairs, significant traffic delays and, in the worst case, unsafe operating conditions. These problems are equally applicable to bridges, especially when concrete is critical to maintaining structural integrity. The sooner problems are discovered, the faster they can be fixed, which saves time and money with minimal delays. Ultimately, this helps ensure the safety of the roads themselves for those who ride them.

However, detecting these malfunctions can be a very difficult task to do manually, especially since cracks that form early are difficult to detect with the naked eye. Predicting where problems may occur ahead of time so that appropriate measures can be taken in advance is also a serious problem. Fortunately, technology is here to help solve these problems.

Bridge building

Built over 20 years ago, the Great Belt Bridge is a suspension bridge connecting the Danish islands of Zealand and Funen. The holding company Sund & Bælt, which is responsible for the maintenance of the bridge, worked with Microsoft to implement an innovative solution that combines the flexibility of drones and the power of artificial intelligence (AI).

UAVs are used to fly around the bridge and create thousands of images of the concrete structure - a method that is much safer and faster than instructing the worker to hang 200 meters above the ground to take pictures manually. Instead, the experience and knowledge of these workers is used to train a machine learning algorithm that can automatically detect cracks in concrete surfaces after uploading photos to the Microsoft Azure cloud. After the AI ​​creates a list of areas of concern, the same experts help in choosing the areas that need maintenance and repair.



Left unattended, these cracks can become quite large and expose the steel frame of the bridge itself. If steel begins to rust, then the strength will decrease so much that rebuilding will be the only way out. “Concrete doesn't just collapse overnight — it's a slow process. Therefore, being able to pre-detect and predict potential damage points is extremely useful, ”says Mikkel Hemmingsen, CEO Sund & Bælt.

The end result is a process that enhances safety, saving time and money. In addition, the technology also allowed the company to use its knowledge in new bridge structures. Further training on the algorithm is also planned, applying the same method to the Little Belt Bridge, Vejlefjord Bridge, and resund Bridge.

“Our main task was to create a solution to maintain and improve efficiency, but we quickly realized that the more we use the solution, the better it becomes. This gave us an incentive to share access to technology with others, ”says Hemmingsen.

From bridges to roads

Multinational construction company BAM Infra Nederland and OrangeNXT, a leading software integrator, have developed a system using Microsoft Azure, machine learning and AI to train algorithms that can accurately detect and classify various types of damage on paved surfaces.

In the past, BAM sent drivers in cars equipped with cameras to take photos and videos of road surfaces. Then the inspectors looked at the materials to identify damaged areas, mark them and make a plan for their elimination. “The process was time-consuming, expensive and tedious,” said Kitting Lee, Director of Commerce and Innovation at BAM Infra Nederland. “We needed a smarter decision.”

The new solution allows cars equipped with 360-degree cameras to take videos from all sides and upload them to the Azure cloud on their own, where AI-based algorithms automatically note any cause for concern. These images also capture geospatial data, which allows inspectors to accurately track them to their real location. This increases the speed, quality, efficiency and accuracy of visual road checks, providing the ability to predict asphalt maintenance time and at the same time reducing costs - freeing up the time of inspectors so that they can focus their knowledge where it is really needed.

“Most roads,” Lee continues, “are checked only once a year. We knew that if we could carry out more frequent inspections, we could prevent the transformation of small defects into large holes, which would improve traffic safety, provide preventive maintenance and reduce the number of emergency repairs that blocked roads and caused traffic jams. "

In addition to the obvious increase in efficiency over time and costs, the new system also helps improve the mood of employees, while attracting new talents. Instead of spending hours looking at footage with undamaged roads, inspectors can now focus exclusively on pieces that need attention, which leads to faster repairs and higher job satisfaction.

Given the success of the new asphalt testing solution, OrangeNXT and BAM are also considering selling it as a Saas solution to other countries and for other purposes, opening up new business opportunities while ensuring their competitiveness.

From bridges and roads to everything in between, these technologies help ensure that the paths we travel are as safe as possible.