# Optimizing Infrastructure Monitoring with ChangeAerial - Event: Xpress ’24 - Published: June 3, 2024 - Length: 9:45 - Watch: https://nestgen.org/library/watch/LdrgTZUgVfc/ > The transcript of a recorded NestGen session, for reading or for use with an AI tool. It comes from the video's > captions; automatic captions can mishear names and terms. ## About this session The sixth episode of #NestGen Xpress was joined by Pete Coulter, Chief Operating Officer of ChangeAerial, a company focusing on drone-based infrastructure monitoring and lifecycle performance optimization. This episode dives deep into the world of infrastructure monitoring, where Pete addresses current concerns and challenges faced by the industry and remote drone operators and how Change Aerial software enables dock operations. Furthermore, Pete discusses what advancements or developments he foresees in automated drone-based infrastructure monitoring, and the challenges we need to overcome as an industry. To know more, visit: ===== Subscribe to FlytBase YouTube - Follow Us On - LinkedIn - Facebook - Twitter - Instagram - ## Transcript **[0:10]** hello and welcome back to another very exciting episode of the nesten express season 3 we are live from San Diego Convention Center this is day three the last leg of the show and for this NestGen Express episode I'm joined by Pete couter Chief Operating Officer of change areel a company focusing on drone based infrastructure monitoring and life cycle performance optimization welcome to the show Pete thank you how has the show been so far for you well you know it's been really good I spent a lot of time in the expo hall just seeing a lot of new equipment talking to the experts here getting great information and uh really really happy to be here so let's just get quickly started with an introduction of yourself and a little about change area okay so yeah I'm Pete Coulter um I've been a researcher at San Diego State University just right down the street here and uh you know we spent like 20 years working with **[1:11]** NASA and NSF and DHS funding um refining a technique that we call repeat station Imaging which means getting a camera back to the same position over time and being able to quickly align images on a frame by frame basis and then compare those for change analysis and um we've taken that technology we've haded it and we've put it into a company change aeriel and we're looking forward to bringing that out the world and getting people using it interesting Inception so uh you mentioned that focusing a lot on infrastructure monitoring so what are the current challenges or the current state of affairs speed like how are infrastructure monitoring are happening today or what are the major concerns around it so if you think about you know Bridges or electric utility infrastructure um you know there's kind of a a history and a sequence of technologies that have been improving things over the years people used to climb and get in dangerous positions and look at things and **[2:14]** then you know there's helicopters and drones taking pictures you know just in a manual way visual review of those very time intensive uh these days it's more drone activity with machine learning algorithms that are trying to identify defects and problems that need to be addressed but we think that that evolution is not quite at the end yet the next step instead of occasional uh inspection every several years and you complete it and you say okay we're done let's pack up we'll move back in a while we think that it's long-term monitoring so it's more frequent image collection and it's using those images smartly uh considering the time aspect to identify defects now right definitely I agree to what you mentioned especially there's a very important need to increase the frequency and I think that's where drone in a box systems comes in a big way so let's discuss doc so how is change are change aial software enabling doc operations so change aiel has two software packages one is a **[3:18]** flight planner and one is an analytics package and so the Drone docks and flight based software Falls really right in between those two right so you know we can do a flight pin coin flight planning so it's oblique pewing images with pincoin targeting on things that are features are of interest uh you know defining view angles and Pitch angles and then from that and defining exact ground sampling distance or space or resolution uh generate camera stations where the Drone will precisely go to a point turn the point at the Target get that exact image that you wanted and then from that point the that you know that you've got that automated flight plan you can repeat that over time so it's an ideal application for a drone dock where a drone's going to fly out repeat this very accurate pattern and it's going to do it over time collecting these images that essentially it acts as like a fixed uh mounted camera **[4:21]** where the image frames can stack over time like video frames okay so that's what our analytic software does is it aligns the image and then does a comparison with AI based anomaly detection indicates where those has have occurred and then we've got a viewer that an analyst can go in and look and see just where those detection results were uh validate whether it's real or not or of concern and from there that can trigger a work action and another interesting thing is that you know today's machine learning algorithms they're really trained and focused on certain things right hundreds of images are thrown in and said this is what you know a broken insulator of this type and of this color looks like and it takes all this training and the algorithm when it's trained enough it'll find that particular issue but it's not finding anything that's not trained core and that's what the change aial software can do we can **[5:24]** detect anomalies um very accurately and then that information like I said can trigger work action or it can also be used to refine the machine learning algorithms so our approach really you know we understand what something's supposed to look like and what shadow change and noise signals are and we can pick out those precise differences that are got it and uh you because you mentioned AI do you think change generated platforms also does a lot of predictive maintenance or proactive uh maintenance yes yes the the approach is really ideally suited for that because you're collecting this temporal stack of data it's analysis ready data that's very consistent it's it's a curated data set that can then be used for Predictive Analytics looking into the future and if you had to talk about one key uh differentiator between all the other products or among all the other products that exist in the market today what **[6:28]** would that be I think it really is this this ability to detect anything this any problem that's not specifically trained on got it so what kind of assets or infrastructure do you think change areas products is best suited for monitoring so the algorithm is is robust is General um it's been used for looking at Bridge damage road damage uh utility infrastructure damage so um it does have a a diverse our beached marker right now is electric utility got it um you know Bridge is a great another great case where you know you're looking if a crack has grown and and trying to monitor frequently if there's been changes um another application that we're working with is micro grids for you have you know self-contained uh electrical system with solar panels batteries all the wiring but it's all self-contained and when you get one of those started you actually want to be able to uh monitor that **[7:31]** with thermal imagery to be able to make sure that everything's functioning properly properly there's not uh you know thermal runaway or anything like that that's a concern so um you know this approach fors very well for that as well got it uh so I think we can end this with one last question which is what are the advancements or developments do you foresee with automated drone based infrastructure monitoring and what challenges do you think as an industry we need to overcome where are we headed so you know I think there is uh just going to be increased automation um you know in some cases regulations do need to be adjusted and and you know it it's being worked on but it's uh you know I think it comes down to you know systems like FlytBase where there there's an end to end process right you don't just have the flight planner you don't just have the dock you don't just have analytic software on the other end you've got something that combine those and feed the data through and when **[8:34]** you've got that sort of automated solution plus the ability with regulation Beyond visual side all that stuff to actually do the work then that's when impressive things will happen in the industry yeah I I agree the there's a need for having a more unified solution rather than having fragmented systems because a lot of systems today are closed and that also reminds me so is change aial as well as an agnos platform or is it just specific for a particular platform so yeah change aial is uh is data agnostic and drone agnostic you know are flight plans can be written out to the CSV file and you know brought into a flight plan or it's just an XYZ position for the Drone to go to no pitch and asmo angle C and the imagery can be really the sort of imagery um RGB cameras thermal cameras you know different arrays uh we can bring all that into the process and and do the change in got it thank you so much Pete for joining in today we are again live from San Diego and we'll come back with **[9:38]** another exciting episode so stay tuned thank you [Music]