 Thanks a lot. Can you hear me? It's a really funny thing. Mostly I'm an old white man, but now I'm a dinosaur, so it's really dead But now one very good thing is This year I have my 20 year anniversary of using drones civil drones I start in 2003 as one of the very first though, and I'm very glad that About the development how much we have achieved and where we are and so I have the big pleasure to tell you about the ice based inspection system for automatic control of power of pipelines as you know, we have around in Germany only 120,000 kilometer of pipelines and with all these problems at the moment There is really issue that we control and monitor them The whole project was funded within the Copernicus content, which is really a very good thing of the European and Union now to our company. What do we do? We have very high precision data for We have very high precision data with drones and other remote sensing and a very good things we create workflows really not from the problem to the solution and we are And we are developing drones Okay, okay So now we go back to the company right up Copernicus I said what what we are doing We providing solutions every solution with a geospatial context. We find solutions before we Collect the data we analyze we make the in which extraction and we trigger action We are we have a company building very large drones up to 23 meters And at the moment we are we have grown we started four years ago, and now we have 18 people and We are have different segments smart farming there in vineyards We are very big in smart forestry because in Germany. We have a big problem with forest with bark beetle and we have the robotics for the pipeline inspection and power line inspection and Of course, we are developing very powerful weed tolls and we make sensor integration The research projects we have the business normal business cases standard workflows But I'm coming a little bit from the developing side And so we are erosion and landslides with big problem in Georgia in the high Caucasus regions then again With pipeline inspection power line inspection. That's very interesting. You can see in our house call you can see here in our Boot we have a very nice hydrogen fuel cell veto with about six to eight hours flight time and eight kilogram payload so really very interesting stuff and Then we make as well we have the forestry and For a 6g the one which is after 5g we are setting standards and developing for drone applications And that will be really a revolution because with the sightling we have very big data rates. We can For here are our drones fleets from the normal copters to the normal vetoes Then the vetoes we have as well for the how's it called? So The hydrogen fuel cell up to the antaris with 40 hours of flight time and this gives a lot of options for Monitoring forest fire and things in that The momabee project this pipeline inspection. It's us. We have spleen lab This is artificial intelligence one of the leading in East Germany Then we have a university and we have dick bro dick bro dick bro is a It's a company. They make gis system for power companies. I think 60% of the European Pipelines are in the software and that's really quite good. So our information goes in the gis And so we have a seamless integration in the workflows How is the power pipeline inspection done to till now? we have the power lines and The people want to know whether they are working or not and so They go with helicopters every month the fly the pyre pipelines people sitting there Monitoring changes things like that Very inefficient very objective, you know in objective some people see things not a helicopter hour It's about three and a half thousand euro and we don't have really not Valid data Step and we our we itself we want to fly with drones. We record the data. We make images We made image processing Information extraction and then we can really see changes. We can see hot spots everything like that It's really amazing that in the year 2023 most of the pipelines people look here. There is something or not. It's amazing how Not very state-of-the-art they are And what do we do we do automatically testing and recording of the linear infrastructures and we want to be Objective we want to be fast. We want to be cost-effective We want to keep databases and we want to understand all the processes and All these is done and we want to make a standardized workflow. We have we are quite strong in Africa And we have a lot of contracts now for power line inspection and as soon as we have the The new technique we can upscale it to other countries and other things So we have different work patterns on one hand. We get the information we get the labeling we get the detection of the failures and so on and Then we put it in the we integrate it in the G is for of the Of the power line in structure so drones We getting the data Artificial intelligence extracted data and then integration in the GIS So the pipelines are a little bit different. We started with the power lines a long years ago But pipelines have a problem. They are only 1% of the time over the ground one 95 and So for these which are over the ground the moving hazards that people run against it are 5% and Environmental impact that somebody falls on it or they get washed under or something are 95% and then for the 99% They are under the ground. We have no ground penetrating radar. So we see Mobile hazards. That's very little In a very big excavating machine who runs over it But mostly they are visible changes that people excavate things. They drill wells. There are Movements or something in the ground so We found we made catalogs of things which happened which are we are interested. I think it's about three four hundred criteria Just one. It's a very important is this soil removal stuff that people just Excavated another thing is that the people drill well for welds for their garden or for the swimming pools or We see vegetation that the vegetation is changing So when the trees are dying this can be an indication that the pipeline is broken or for other things as well in in winter for water pipelines when Frost because the water is warmer you see very very characteristic Characteristic patterns And now we're doing automatically Analyzes and we want to do it on board. We have an DVI card strike So when there is something it will be detected it will be extracted and it will be compared with the Well, it's labeled and then and this is a very important thing. We have probabilities whether how pro with which probability this is it and when we are don't have the when we don't have the probability and We fly on the one wave up Normally, and then we fly the way back from a digger angel in a lower in a low in a lower Height and there's this then we have every object of different angles and different things. This helps a lot for the object and Recognition and this is the capturing process We have a video camera arm and then automatically High-resolution images well taken and they are linked to it. So and the next one As soon as they come now we see here the marker posts the staircases things like that Is it allowed? It's not is there some changes then in the next image bum We get as well high-resolution from the way back from thing we make 3d models as well for the object cars You can detect that with but easily we see the fear vehicles in the in the way like that Next one and as well. We are trying to do sensor fusion. We have multi-spec cameras to see vegetation we have leader for getting the objects and Here is so we get suspicious new planting So when some farmers try to build up new trees with deep Roots, they may be a threat or here. This is a washing This is one of the biggest issues we have in Germany that that the soil is washed away So I want to make it short because you're all the beer is waiting We have a boot in the room 5.2 and I really would discuss a lot because in the whole field. There's so much bullshit on the road I know some providers who do in it and say they have AI and I know that they sent the video streams to India and the People there there are hundreds of people just visual interpretation, which is not bad. It's a good method But it's not AI and but what what is our lesson learned? so the inspection of under Pipelines underground are completely different like the Just like the power lines. So a lot of people which have power lines want to go to Pipelines and they face issues This and the biggest challenge is that we detect the changes on the soil surface Which are indicators of broken pipelines or damages in the pipelines? the biggest thing for the AI and that we have really so much bad experience was the training and Which is AI is very perfect when you have proper light conditions when you have the geometry But in nature everything you have a lot of degrees of freedom when there's a little bit snow on there when there are leaves there When there's litter laying when there is some Different even colors and this really creates a lot of issues And we are now much further than three years ago when we started but we are long not We have a long way to go before we get really this absolutely 1995 percent we need and Which is a very good idea what we did is that we that we work with the detection that we get the Probabilities and so we say this is quite sure so we don't have to do and then the other thing where we don't know Then we can go fly deeper or some some operator has to look at it and with these we can reduce the Quite cost and space the work of the labors by far and which was very nice Experience the long-range drone we have that's really a very good tool because on the one hand we have 4 to 6 hours flight time and other on the other time We can interrupt all the time and can hover and we have still a long off For detailed inspection and we're still a long way to go. I thank you very much. I think this was the Yeah If you have any questions, I'm very happy to discuss it at the boot or with a beer or whatever And I really would like when somebody has good ideas when somebody is further than us Please approach me for me that I'm working out 20 years with drones and I'm very very frustrated That things I have did I did 2008 in Australia like irrigation control with drones Still it's not standard. Yeah all this bug beetle detection. It's possible But it's but the the problem is the companies are too small And so they don't have the time to make proper testing and the big companies or like DLR or something like that They are not efficient. They have so much. They are not moving and I think very much when we are getting together The people and we've been in the one case can these do this then we get much further for the benefit of all Thanks a lot