# Patrolling Dutch Rail Corridors Against Graffiti and Trespass with Autonomous Docked Drones - Event: NestGen ’26 - Published: October 2, 2026 - Length: 31:09 - Watch: https://nestgen.org/library/watch/H_988Peyun0/ > 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 Rail networks are difficult to secure with fixed cameras and manual patrols alone - long, exposed corridors make graffiti, trespass, and vandalism hard to catch in time to act on. Autonomous drones are becoming part of how security teams extend consistent coverage across these hard-to-monitor stretches. In this NestGen session, LUGN Security will share how autonomous docked drones support rail corridor patrols against graffiti and trespass in the Netherlands, and what security teams should think through when deploying drone response across long, linear infrastructure. ## Transcript **[0:00]** Welcome to this session of NextGen folks. I have Alex here with me and first of all like feel free to drop in which country you're joining us from. Help us understand I know this is the email address so I'm sure there's going to be like people from various different countries. Introduce yourselves in the chat. Today we are talking about a very unique use case in this session about specifically so how Alex has essentially built a drone program for patrolling Dutch rail corridors but not for the use case you may think for. It's actually against graffiti and trespass with autonomous dog drones which is a pretty unique use case. And if you think about it essentially graffiti damage may not seem like something huge but it can cost like tens of millions of dollars for for the government and it's very interesting to understand like how actually that much ROI is being generated with dog drones for this use case. And with me **[1:03]** today I have Alex over here. So Alex is the founder and CEO of LUG Security and he has also been the founder of LUG Security of just give me a second I think my screen is glitching a little bit. Okay. Right. Right. So I was talking I was introducing Alex. Alex is also apart from Loon he's also the director of Nemo Dog and Event Security which provides physical security with dog handlers and like that experience of his in security operations kind of shapes how Loon designs their drone programs as well, as part of their existing security operations, not as a separate like gadget. **[2:04]** Loon Security specifically is a licensed BVLOS drone operator. They're based in Heerlen, Netherlands, and they're running drone operations for third parties across critical infrastructure industry, logistics, and large-scale sites. Just before the session, Alex was talking to me about like some new deployments they're doing, so which is super exciting. So, Loon owns the full stack hardware, authorizations, pilots, systems, airspace coordination, operational control with their own 24/7 staff control center, and an in-house dev team as well. They fly more than 1,000 flights per week with one pilot currently handling four aircrafts at a time, which is like incredible feat in terms of drone autonomy. And their role is to detect, verify, and escalate. The drones kind of the drones kind of sit as an added layer on top of their client's existing security setup, and the client's own control room decides on the follow-up on any incident. So, speaking of the Dutch Railway **[3:09]** deployment, their work with Dutch Railways flying BVLOS at a railyard with a fully automated chain from alarm to secure video stream landing straight in the client's own video management system. I think that's all like really, really interesting to hear about. So, I'm just going to take like 1 minute to do some housekeeping over here. If you find Alex's session valuable, I'm pretty sure people in your network and your colleagues are also going to find it valuable. So, please feel free to take a screenshot of the session at any point, share the the screenshot on your LinkedIn. You can feel free to tag NestGen and tag Alex and Loon Security as well. That is going to be uh of course really useful for people who are looking to exchange more inside. Uh secondly, we are going to be publishing the this whole session is going to be recorded and we are going to be publishing it on NextGen's YouTube as well. So, if um you know, you want to take some notes, come back to the session, uh you can feel free to do so on uh our **[4:11]** YouTube. Uh and without further ado, I think I will pass on the stage to Alex. Welcome to the NextGen stage, Alex. >> Thank you very much for the for the vast [snorts] introduction. Uh I will share my screen now and I will go straight ahead to >> Right. >> to the presentation. So, uh everybody welcome. Uh my presentation is about patrolling the Dutch rail corridor against graffiti and trespassing with autonomous dog drones. If there are any questions, feel free uh uh ask them. I will try to give an answer right away and and uh yeah, uh yeah, let's start. >> Right. >> So, uh first uh I will I will talk about the the operation, what the rail work on a rail yard actually is. And and um yeah, **[5:14]** for security guards, it's a it's a it's a hard um an environment to to work because um uh there is a lot of safety and and so it's uh it's not easy to to walk from from the A to B. Uh hey, you and there's a lot of traffic and a lot of things you have to keep in mind. Um uh yeah, for the graffiti, um the the objective for us is to disrupt the the graffiti the the the piece has to be broken. And and because it's very hard to an arrest a graffiti sprayers. So what we call a whole whole whole car so a whole piece of a train takes about 1 and 1/2 to 3 hours to to spray and it **[6:14]** takes 30 to 30 to 40 cans and normally a crew are three or four people to to bring that time of 3 hours back within a 1 hour. So every interruption that does not happen that's that's that's a win. And the the the success of the whole operation is measured by preventing damage. Now why it's almost impossible like I said all the rail yards are very very large there is constant motion freight international services. For example the Dutch railway track is about between 6 and 7,000 to kilometers and it's one of the busiest I think the second busiest after Switzerland of the European railway track. And **[7:19]** there are not only trains which are which are staged on the yard which have to be cleaned or serviced but there are also active trains who cross the yard. Yeah [snorts] and the guards like you see in the picture they also wear high visibility clothing so it's for the graffiti sprayers very easy from a long long distance to detect our our guards. >> [snorts] >> Yeah, and arrests do sometimes happen, but that's more a co- coincidence really by design. Then the drone. About 3 years ago, we started to do some proof of concepts with with with drones. And and the drone is hardly visible for the graffiti sprayers. The the guard is, but they don't have a clue where the drone is, especially when you **[8:23]** keep a certain height and a certain distance. And and Yeah, like like there is in the bullet, the asymmetry makes the makes the operation ineffective. This is a picture from from from a railyard and and what we do is we are do something we call heat heat mapping. It's like It's like you do with you can see on Strava or other running apps. The busier certain track is, the more it highlights or is or it is red. And the same thing we do with with with rail tracks and graffiti. So, we register or try to register the the square the square meters of graffiti on certain tracks. We try to find **[9:24]** out which days, which hours, and that is how we build a heat map based on the on the historical incident data. And this is an example. And and it gives an insight instead of of of doing things by by your gut, we have pretty good understanding of what is happening because most graffiti sprayers are not kids from 16, but they are uh uh upper middle class plus 30. Most of them are well educated, so they have a job, they have a family. So, it's it's so you can see a trend in when they have their day days off. Now, how do **[10:25]** do our does our operation run? Uh you just see uh see a FlytBase a flink board and and we have detected the the the or sorry, we have connected the parameter protection cameras, all other kind of stuff. It's it's connected to to a platform and from there we we we uh from there we build link flinks. So, it's very easy for our pilots to see if there is a pop-up and and our pilot sends out the the drone by just simple pressing the go drone button. And and **[11:25]** yeah, the the the drone is within 2 or 3 minutes. It is present It is present at at the at the spot where there was a a detection. And besides incident driven we also fly the patrols. So we have a we are using the the the FlytBase scheduler and we and we pro program our at at random patrols. Um if we detect and want to want to do an intervention, we use the the speaker and the spotlight on the drone and and with the with the railyard control room we we co coordinate **[12:26]** the whole action. They get the live stream from the any event. And in the Netherlands there is a protocol it calls it called live view where we can also share the live stream live stream with the Dutch police. Now this is this is a view from our our control room. Who decides what? Now the live imagery goes directly to the client's owned owned security control room where there is a security specialist. The data they they they handle the the response and decide what is recorded and when the mission ends. Um we we don't build something we we don't call privacy by design. So, **[13:29]** it's a client's choice after landing where the footage is written. If it's in in in our the client's own VMS or if it stays on our servers. As we speak, we are we are transferring from an Amazon European server to a Dutch own server. So, yeah, actually we fly and operate the payload, the cameras, the lightning, the speaker, and and do it by request of the security control room. Yeah, automation is not autonomy. When an alarm arrives in the flight management platform with an audible an alert, we use our pre-programmed missions and we fly them automatically. **[14:31]** Dutch law requires that the drone operator is physical and can has start start the mission, but more is capable of taking over the whole mission at any time. That's mandatory. That's not something but but a choice. We have a one one-to-many permits. So, and we use that mostly for the patrols. They are all flown robotized and at random with full attention for the site, but mainly monitored by a AI which is in the loop. We don't use the vehicles and detection plus some of our own own models. Um yeah, and the our promise to the client is that within 2 or 3 minutes **[15:32]** they have a life a life uh life image, so full situational and awareness. Yeah, [snorts] then then the case, what it saves and what it scales, these are official numbers from from from Dutch star railway. You can find them on their website. The annual costs for uh graffiti cleaning are about 10 to 12 million. The 12 million is also based on they take trains out of the over of the operation. They have to put in a new train, so it's a lot of hassle. There are about rough actually 40 sites in the Netherlands where trains are stationed, so the average damage is about 300,000. If you If you um so that it's so that's 25,000 **[16:35]** in a month, and we provide drone as a service, which means our clients don't have to buy any system, so we do we have we do we do our the all the holder of the permits, we do the maintenance, we do the hardware, 24/7 pilots in our own control room. We We are in charge of the of the airspace, and it depends on how the how many hours a day it's about six the amount is between six and seven thousand euros monthly. So So And then What it saves on the sides where dock docks are deployed, the average **[17:38]** reduction of graffiti is 90%. So Yeah. This yeah, this is not a very hard um calculation to make that that the that the costs of the docks are are are a fraction of the of the actual graffiti damage. So like I early stated, it's a one to four one to many so one of our pilots supervises four aircraft at the same time across different sites and different clients. At this moment we make more than a thousand flights in a week across roughly 10 locations and this is a mixture of patrols and alarm response together. **[18:41]** And yeah, like I like I said earlier, you can buy a docking station. But you cannot buy the whole organization underneath it. That has to be built and you have to gain an experience and that's what we did in the past six six years. So we are building a operational we are building operations manual. In in in the there's the in the Yazza, we have uh there are mandatory roles like the accountable manager, a safety manager, flight operations manager, maintenance manager, training manager, security manager. And yeah, and and operation and safety that is a hard a hard split. Uh so people who are involved with safety cannot do operational tasks. **[19:44]** Uh we have at this moment nine fully trained remote pilots on an STS-1 STS-2 level with also an endorsements on overflight BVLOS and night. And of course everything we have the right to an insurance and a 24/7 audit trail. So yeah, I think that we were right on top of it and and because we think that deploying drones in security is more than selling a drone and and and training a pilot. If I sell you a drone and train only a pilot, actually I'm not selling a solution. I'm selling in a problem because afterwards you will find it's a lot lot more. **[20:44]** So yeah, I've seen that I've gone a little bit fast maybe through the presentation. Uh I thank everybody for their an attendance and and I'm going to stop the presentation now and we can go back and and maybe I can answer some of the questions or >> Absolutely, Alex. Thank you so much for this incredible insightful presentation. I think the one-to-many ops and especially like how you are able to prove such ROI and at such a grand scale. I think that was all really insightful and a lot of people like while you were presenting a lot of people actually started like posting questions about the operations and like how you're running the whole program. So, I'm just going to like start sharing the questions one by one on the screen. For you. So, I think Robert Peacock has asked at what altitude do you tend to **[21:46]** operate at in order to avoid being heard and detected by observed intruders because you mentioned that like the whole point of like flying a drone is that it shouldn't be visible or recognizable for the intruders. >> Yes. Thank you, Robert, for the question. Like we run two types of of the operation when we have a detection and and that's by choice of the control room. One operation is to to the to we don't call it stealth mode. So, we don't want to be detected just monitoring what is happening. The average height is about is between 60 and 70 m because that's when the split view from the thermal and the regular camera is more is the most ineffective. If you go a higher than than than than than the thermal **[22:50]** yeah, doesn't work that well and and we keep a distance about 100 to 150 m and and and there's also some noise, of course, on the train yard, so they they don't see the drone. And when we do an intervention operation at for example, then we our maximum height is 30 m, 35, and we come close, say, 30, 40 m, put the spotlights on, play play audios. And what's important when you're working in areas like like the rail yard, you we we we try to push them out of the rail yard, so not make them run further into the rail yard. So, hopefully hopefully this answered your question, Robert. **[23:50]** >> Uh thanks, Alex. Another one up here. Rail corridors are long and linear. How many dogs does it take to cover a meaningful stretch? And what is the gap between each dog's coverage area? >> Uh um yeah, actually we we can do it with a one one dog. So, um because we don't really guard the rail corridors, we don't guard the rail yard. So, the average yard is maybe 3 2 3 km long. So, when you choose a strategic the position that the perimeter is within 2 minutes. >> Okay. >> Within 2 minutes. >> Got it. Um pulling up another one here. Graffiti and trespass are two very different problems. One need evidence collection after the fact, and the other needs real-time detection and response. Does the drone handle both, or is it optimized for one? >> No, no, it can handle **[24:55]** uh, um, handle So, that's uh, we and of course we uh, act on different scenarios. Uh, for example, in in the daytime when the graffiti and the trespassing are not really happening, we for example, uh, scan scan for example solar panels or we do a do do a fence detection which is during uh, bad lighting more difficult. >> Yeah. >> So, you you you see the more the longer a drone dock operates at a site, more more tasks are added. >> Absolutely. Uh, so Neil is asking Alex, uh, have you considered importing and the spraying area locations and material stock directly into the FlytBase dashboard? This would consolidate all operational data into a single view and enable faster response time. >> Uh, sorry, if I have considered on an importing and then? >> The spraying area **[26:01]** locations and material stock directly into the dashboard. >> Uh, no, yeah, we we are we are building our own dashboard where we gather all that kind of data also live live data from from the trains where they are. So, um, and and and sharing that um, that that data with the guards so that they don't walk to tracks where there are no trains or better that they walk to tracks where there are trains from which we know that they are very popular and amongst the graffiti sprayers. So, actually it's not it's not about collecting data, but it's building about the information. >> Absolutely. Um Another one here, I think this one is very interesting. Rail corridors would cross **[27:02]** multiple jurisdictions. Do you need separate approvals for each municipality or does one national authorization cover the full route? >> Uh the answer is the is is getting lost. Our national um uh >> BVLOS approval. >> Yeah, the uh yeah, the national BVLOS and approval covers all of it. So, we only had to deal with the owner of the railyard. So, the so the railway and company. >> Got it. >> And and of course with the airspace, the airspace uh uh was was guarded by the Dutch military or authority, but that was rather simple. >> Mhm. >> Yeah. >> Um I have another one here um from Hasco who's asking, do you fly 24/7? If so, how do you handle the obligation to publish NOTAMs and report BVLOS flights outside of UDP? >> Uh in the Netherlands, we **[28:05]** publish a NOTAM and it has an has a validity of three three months. So, we have to publish every three months a new NOTAM and um our NOTAM says that we are 24/7 flying. So, we don't have to put one out for every flight. So, we don't even even have to have to report that we are flying. So, yeah, we got yeah, that's that's that's what we have earned in the last last years. Yeah. >> You've got a quite a few questions here. Um good sign of like I mean, the presentation was quite engaging. Uh so this person is asking we run security for a transport network and our challenge is the sheer length of what we need to cover. Is this problem cheaper than CCTV along the corridor or is is it solving a different problem? >> Uh no, I think it's um it's the cheaper in the long run. **[29:07]** Uh for example, um we are Yeah, they are also now testing with the fixed-wing uh drones. Uh so uh drones who can do uh longer stretches uh for a hundreds of kilometers uh and they are more used for the an inspection purposes. And >> [clears throat] >> uh we we focused on on the yards and and and workshops and other what's what's called more critical sites. So, I think that in the long run, it's much cheaper than CCTV TV. >> That makes sense. >> Uh absolutely. >> So we are already running out of time and there are other sessions in the line, but I can share like one last question here. Uh Tom here, I might have missed this. You're flying over people. Are you required to use means of ground risk mitigation? **[30:07]** >> Uh yes, we are flying on the Aurora 2.0 and Aurora 2.5. So we have to have to comply with that. Yes. >> All right. I think that mostly answers all of the other questions. I know that I may have missed like a couple of questions uh here, but we are really short on time now and overrunning already by a minute. Uh so I am going to drop Alex's LinkedIn in the chat box. Uh anyone who has been who has felt inspired by Alex's and Loon's work, please feel free to go and connect with him on LinkedIn, understand uh his insights in depth, and I'm sure like Alex is is always of great help. So, I'm sure he's going to be right around the corner. And thank you so much once again, Alex, for this incredible session at NestGen. Of course, like it's a privilege to have you on this stage. And really, I I even I learned a lot in today's session. And I'll see you. >> Thank you all. Thank you. Bye-bye. >> Bye. Bye, Alex.