 Okay, so today we will talk about the three things AWS green grass how to build your own drone 3d print your drone plan and 3d print your drone and then how to use or it is recap and drone and AWS green grass to Aerial photograph the photograph of a site and build a 3d model So let's go agenda So let's talk about AWS green grass first. So AWS green grass is basically for edge computing So people who know IOT IOT needs IOT is all about connecting your sensors so many of sensors are generating a lot of data and Some critical decision has to be taken at the site where at the point that the data is generated So if you are you if you are sending the data to AWS cloud and taking the decision by the time the data comes back Probably the event of the lab for example the Tesla car if you're going to Tesla and there's an accident going to happen You have to take the decision at the point of where the Incidents is happening. So you can't send the data to cloud and take a decision and come back whether to apply the break or not So AWS green grass is a solution for that. It's called edge computing So what database green grass gives you is it gives you an SDK to take that SDK and Flash it on to your hardware like Raspberry Pi or something like that And then you can take all the Cloud decisions that you are supposed to make on AWS cloud. You can actually do it on your local On your local Raspberry Pi or in your hardware So basically what it does is you have a lot of sense here and this is our typical IOT world You have sensors you send the data to the cloud and do then cloud Do they you store the data and then you take a based on the data that's coming in you actually take some actions like say You might send an email SMS alert or something like that, but that is actually Not very useful and most of the critical situations. So where what we need to do is we need to do process the data This shows where the data is generated. We need to process that. So as part of that we have AWS green grass. So what are the value of processing the data source loss of physics loss of physics is that Anything so that has to travel it takes some time So to reduce the latency from sending the data from the source to the cloud You need to use something that can process the data right at the source So that is the loss of physics because of lots of physics. We need to ensure that the data process the data As soon as possible The second law is he says that loss of economics a typical AWS IOT scenario what happens is the Sensors are generating data every second or every microsecond like every 500 milliseconds And all the data might not be useful to you You might want to process only certain bits of critical information So if you are able to actually process only the critical information and send only the important information to the cloud It actually saves you money. So loss of economics dictate that you need to actually have process only the critical information that is required in a typical say if you take an Boeing 747 which can fly for if it flies for four hours exactly Then that's one terabytes of information and all those one terabytes of information You don't have to actually pump in it to be your database You might need only the critical information that is required So you can actually process at the source and then only sort the local things And finally the laws of the land laws of the land says they might be privacy issue or some Government regulations might be there. So based on that it might be required that the data has to be within that Source it doesn't have to go out or something like that. So based on laws of the land You might have to do a process the data at the source. So all this requirement is actually kind of satisfied by AWS greengrass So if you see the structure here earlier You saw that there were sensors sensors are sending the data to the cloud and Process of the data is to be processed the cloud and then action is to be absent happening at the cloud now whatever is you have The sensors the sensors will send the data to the gateway Which is right at the source and the gateway has AWS greengrass net and AWS greengrass It process that data and can take the action there itself right at the source It doesn't have to actually go go to the cloud That's not a necessary condition that to work for AWS greengrass to work So what are the benefits of using it up is greengrass? Like I said responds to local events quickly There's some event happening and you can actually take a very quick action for the events that are happening on the site then you can operate offline in one of the critical one of the Condition one of the scenarios where IOTs are typically deployed is you are not always guaranteed That they are connected to the internet and your connection can go offline or something like that So you need to operate offline. So AWS greengrass SDK provides you that option So all your data is actually managed by AWS greengrass SDK on your local gateway and whenever the connection is available It will actually send that data back to the cloud So it will do does the syncing and all those things for you Simplified their device programming. So AWS greengrass will say you the same programming that we are all aware of AWS programming API's So you can use the same API's and you don't have to worry about syncing and you don't have to worry whether the internet is there or not or all those things and One of one more other good advantage of AWS greengrass is you can actually have Python lambda functions Running right on your local source. You don't have to go to AWS to run your lambda functions You can actually run it right on your local gateway and Last reduce the cost of IHP picking out already Reduce the cost like I said like a Boeing 747 flying for four hours generate one terabyte of the data If you're sending one terabyte of data to cloud, it's cost you money instead of that If you can actually send only the critical informations that can reduce the cost of your IoT applications So let's we covered the AWS greengrass now. Let's make go to how to make a drone Okay, so Yeah, in this figure, I'm actually I built a drone or my own custom drone I designed I actually work for autodesk, so I took autodesk fusion 360 I designed my own custom drone using a fusion 360 and then I use the hardware called Pixoc. Pixoc is a open-source autopilot hardware for drones It's a you have complete control over that and you can connect all kinds of sensors GPS Your ultrasound sensors and all those things and this is a schematic diagram for how to connect all those things I'll show in a video how actually So what I did was first I went to autodesk fusion 360 I designed a drone and Then I pretty printed that Then I had ordered this hardware assembled hardware and Then why what I did was I mounted a camera on to the drone and then I took the Raspberry Pi and flashed AWS greengrass, so all the information so this has a lot of sensors connected. It has GPS constantly being fed and The battery information a lot of these almost has the 10 different kinds of sensors constantly feeding every 250 milliseconds is it some kind of data is being pushed So all this data can actually store in your greengrass and finally when you have it some kind of connection You can push it back to your AWS cloud for later usage So This is a video of your iPhone iPhone application Apple Apple Maps application if you have seen this There's an option if you go search any US region. You can actually 3d You'll get a 3d map so you can actually have a 3d aerial view of any of the region So there's a New York region that I'm showing here. So you can see that We there's a 3d map so how you can actually build this using your own drone That's what I'm actually going to show you here So you can build your own drone then use Autodesk recap So software to actually create stitch the photos and make this kind of area 3d mapping So there's a small video just to make you know This is 3d printing of the all the parts. There's a various testing that I've been doing it There's a XR copter that I built so Assemble the all the parts tested and I was able to also demo it as part of the make a pair that happened in Singapore So this actually shows how the drone was built or the next part will be mostly covering how we can use recap Reality capture to actually switch all the photos to generate 3d 3d models Yeah, so this is the video so what you saw actually was how we actually designed the Drone and then how did we 3d print and then how did we mounted a camp basically I mounted an Android phone to the drone and then Programmed the drone to fly over certain area and it can actually take take photos at certain intervals Store the datum and I'll send all this information To store it in AWS green grass and whenever net is available It will actually send it to Autodesk week reality capture and autodesk reality capture will do the hard work of sticking the photos and Generating with a 3d model so you can actually at the end result will be having a 3d model of the site that you actually scan using your drone So Apple was using helicopters. They actually flew a copter over the entire new York region to actually generate that entire 3d map now the technology search if these drones will be costing you around $800 and Other AWS green grass. It's actually free for up to three devices three sensors for one year And then you have to use reality capture autodesk reality capture to stick the photos That actually