Imagine you're a farmer in Ukraine. You go out to work on your farm one day, but you never come back. Why? A land mine killed you. Isn't that messed up? Good afternoon, everyone. My name is Navin Agaralc And my name is Zorann. And where UnveilX? A 51c3 nonprofit? UnveilX have developed a machine learning based solution to detect and neutralize land mines that is more efficient and safer than current methods. Landins have killed or injured more than 1 million people in the last 50 years. And the problem with landmines is that they're extremely easy to lay down but really hard and dangerous to detect or remove. Currently, landmines are being laid down 25 times faster than they're being removed. Here's an inconvenient truth. Over 25% of occupied Ukraine has been landmined and the UN estimates that it will take over 100 years and $37 billion to de mind this area using current techniques. The most prevalent mines in Ukraine today are above the surface mines. For example, the Palm three mine. These don't need to be buried underground and are far easier to disperse, simply launched out of a ground launcher. They're small around the size of soda canans and hard to detect yet more lethal. The current solution is mainly ground based de mining. This is time consuming, expensive and risky. Other imaging techniques like using satellites or high flying drones won't work either, since their resolution is far too low. Our solution is to use low cost low flying drones and automate the mine detection using machine learning. Drone operators simply need to record a fly by video onto the SD card on the drone. Then upload the videos to the Cloud. Our model will take over in pinpoint land mines. Like Tesla uses AI to see cars, we use AI to see mines. Currently, we use a $500 DGI mini three drone which is easy to use and can fly for 30 minutes on one charge. For longer operations, operators can use a spare battery. In some cases, even surface landlines can't be detected by drone fly by videos. To solve for this, we've used recently developed long wave infrared cameras to see through the foliage. This helps, since different parts of the electromagnetic spectrum provide different levels of foliage and ground penetration. And we plan to solve the problem of buried landlines over the summer. When our model identifies a land mine, I'll send the cornets to a database. A second drone can then fly to those cornets and drop an object to set off and neutralize the land mine. To create our machine learning model, we first needed to create a dataset, since there was no publicly available data sets. To make our dataset, we had to create scale mark ups of the Palm three land mines and place these in various terrains and make our dataset. To train our model, we use Microsoft Azure machine learning environment because it's currently leading NII and it's startup friendly. We're offering our solution to the Ukraine Deputy Prime Minister's office for about $250 per square mile. This is ten to 100 times less than any other solution out there. And the market potential there would be about $20 million. We plan to scale our company by enabling grassroots organizations with our solution as well. We've contacted multiple combat zone shareholders to validate our solution. In particular, the third special Assault brigade in Ukraine. We've also proposed our solution to the PM WRA a non combat agency for the US Department of State who are working to help clear land mines around the world. For example, they recently cleared 395 acres of land in Bosnia. We also have two key advisors, both who are Ukrainian American and work in Google and Microsoft AI. Our only incremental cost is Azure compute, which is about $1 per hour of video. For Combat Brige and the government, we plan to implement a licensing fee starting at about $100,000. For scaling, we all know that the government isn't enough. We need crowd sourcing to really improve the implementation of landmine detection for all. We've won $5,000 in Microsoft backing, and we're also looking to apply for SBIR and other sources. We need this additional funding for higher quality drones and technology development. The social impact of D mining is huge across so many different countries, not only for lives saved, but for livelihoods. We're unveil X. Let's save lives this safe foot. Very much, Judge. You have 3 minutes. So you mentioned dropping once you identified the land mine dropping something on it detonated, is that what I understood? Uh huh. How are you ensuring that That's okay. Like, how would it work? No, that you are actually that it's you can safely drop an item on to detonate the land mine. Okay. So currently, what we've done so far is we've worked on our model to detect the land mines. What our plan is for our second drone to fly by is that'll drop any object on or near the mine to set it off. And the research we've done on this mine shows that they have seismic sensors that can sense any sort of vibration nearby, which is when they'll go off. And the idea is that if we set it off with any sort of object, it won't be active if somebody comes near it. Thank you. So you mentioned that you reached out to the Ukrainian government and to the State Department. I think you said that you contacted them. What was their response? Well, with the Ukraine government and the third special Assault brigade, all of them are saying that landmines is a huge problem, and they're really desperate for a solution because right now, like the third special Assault brigade, they're on the front lines fighting. They need some way to clear the path in front of them for mine detection, especially so they can move on with their fighting. They're saying that You know, problem of land mines is, you know, huge. This is like a really big solution for them. So what event is, what are they saying about your solution? What what kind of reaction response do they get to your offer? They have a grit exciting. They're really excited about it. Okay. Yeah. So I've seen other technologies that are designed to do this. I've seen pitches farm as well. What is exactly your strategic enabling technology? And how do you intend to defend that from the competition that's out there? Okay. Well, first of all, our product is extremely cheap compared to all other solutions, and it's extremely efficient as well. Our model is highly accurate, and I believe it was around 80% 89% accuracy compared to other solutions, it's a lot faster and more efficient and cheaper. And on top of that, we're using RGB imaging plus thermal imaging, which other companies don't use. So that's kind of what I'm getting too. What is your differentiating technology that you I'll be using the combination of RGB and thermal imaging. Okay. Okay. Have you trained operators to operate the drones and how successful are they at identifying and then destroying the mines. Our job isn't to train operators. Our job is to cloud inferencing. So drone training and drone operating for especially like small drones, like, is very easy. We've gone through the training for them. It doesn't take long at all. So that's mainly on the client institute. Additionally, we've also heard from when we contacted the Ukrainian government that romees are already oftenly used in Ukraine for combat, they're used to sort of drop grenades. And also, many people in Ukraine are already well trained in using these drones. Great. Thank you. Thank you very much, Sean
2024 Diamond Challenge Summit - 6 - UnVeilX
From Andrew Brett June 03, 2024
14 plays
14
0 comments
0
You unliked the media.
The Diamond Challenge provides a unique opportunity for teens to learn about entrepreneurship while putting their ideas into action.
…Read more
Less…
- Tags
- Appears In
Link to Media Page
Loading
Add a comment