How AI and Infrared Technology Are Making Bridge Inspections Safer and Faster

Necati Catbas, Ph.D. (00:05):
I think there's a national need for our infrastructures to perform better. And when we talk about infrastructure, it's not only bridges. There are bridges, buildings, water systems, railroads, railways, all of them are still infrastructure systems and they need help in terms of better understanding their performance and where we can allocate our resources more efficient. So there is a national need, and we have been working on how maybe we can use new technologies to really understand the existing condition of structures.

Marwan Debees, '23Ph.D. PE (00:45):
One of the things Dr. Catbas was always reminding me of is, how can we use the technology to give you information that you really need, that you can utilize, whether it's to assist the conditions of the bridge or to help you in the decision making process. So when we looked at this particular paper, one of the things we were looking at, okay, what do we do right now? What is the status quo? What do we go look at and these bridge inspections and how can this technology help us? We use infrared data to look at the conditions of the bridge. There are multiple components of the bridge. As you know. There is the bridge deck, there is the superstructure or the beams typically, and there is the substructure or the piles. In this case, each component was looked at using a different method.

Necati Catbas, Ph.D. (01:40):
The infrared thermography, I'm going to explain this a little bit more, but it's one of the ND and ND stands for non-destructive evolution. So techniques that can be used and we can use other ND technologies in the framework that behave, but in this case we use infrared. We are not using just infrared cameras and collecting raw data, which we do. And the second thing is we are using a vehicle, and in this case we collaborated with a company from Virginia next USA, and with their vehicle we have the infrared camera and high image cameras and we can collect data at high speed. So it does not interfere with the traffic. So this is an important thing, especially on easy bridges. So if you go and look at a bridge, visually inspected, I mean it's almost like a doctor just looking at you, you look fine, you might be fine or you might not be fine. I mean, you look at the bridge, obviously it's not just looking at, they do certain things, but human visualization has some limitations. It might look okay, but there might be some hidden problems that can, maybe the sensors and the other technologies can tell you. It's almost like the doctor looks at you and then they say, okay, I will going to send you X-ray to MRI, so they can better understand. So we are doing something similar to bridges.

Marwan Debees, '23Ph.D. PE (03:09):
One of the things that I think would be helpful is using the same concept for other types of infrastructure and other types of NDE. So there are two different, not necessarily different, but two areas where we could build on this research, applying it with different NDE methods and then applied it for different infrastructure types. So in this example, the neural network that we've developed is for concrete bridges. However, the same concept could be applied for steel bridges. It could be applied for other concrete buildings like Dr. Sheb was mentioning, for high tower high rise buildings or other infrastructure that requires inspections.

Necati Catbas, Ph.D. (03:55):
Marvin was officially a student of mine, but now he's a doctor. He's a colleague of mine and like other PhD students of mine, I still keep in touch. Now once they graduate, they're my colleagues, they are my friends colleagues. We are very close with most of my, if not all of my PhD students. So we talk to each other and now that he's so close, we always work together, brainstorm together. We'll continue and hopefully we'll jointly guide the new generation of students. And I am asking you, but I know you'll help the new generation in the past, in the future with your past experiences.

Marwan Debees, '23Ph.D. PE (04:39):
So Dr. You already know it's a pleasure and an honor to work and continue working with you.