ArcGIS Apps: Drone2Map for ArcGIS: Bring Drone Imagery into ArcGIS

my name is Tony Mason I am the product manager for drone them out here with me of Cody Benkelman he's a technical product manager with our imagery team in our core development group morning so what we'd like to do is week to spend the rest of this session giving you an overview of drone a map will dive a little bit into the software show you how it works show some of the results from some of the projects that we've been running throughout the beta period and then Cody will dive into a little bit more about how you can take that data and use it throughout the rest of the ArcGIS platform alright so let's go ahead and get started so I'm pleased to announce that drone a map is officially out of beta it is now in its 1.0 release Jona map is currently available through the arcgis marketplace and for those of you who are not familiar with drone to map for arcgis it is a standalone desktop application that leverages your ArcGIS online named user account to streamline the processing of drone data and so we'll we'll go through the session today and show you some of the design elements and things that we put in a drone to map to gear it towards the GIS user okay so first and foremost when we set out to design drone to map we wanted to do it with the GIS user in mind meaning that we wanted to allow you to take affordable drones and turn them into enterprise productivity tools without having to be an expert in aviation or an imagery so it was very very important as we started to build and continue to build drone to map so drone to map for arcgis produces a wide variety of products these include 2d data sets such as ortho mosaics digital surface models you can also create 3d datasets from your 2d still imagery out of coming out of drones these include point clouds and textured meshes we also allow you to work with oblique imagery a little bit differently and I'll cover that in some of the demonstrations here but we have what we call an inspection viewer inside a drone to map and this inspection viewer is all about viewing your oblique imagery in its native resolution interacting with a map so seeing what the camera was pointed at being able to zoom in nice and tight on features draw some annotation and then save those out okay so in order to get these products I just wanted to stress the importance of planning your flight to get the right product type and so when you're setting out to create 2d products we recommend flying in what we call a lawnmower pattern or a grid pattern and this is taking your camera pointing it face down at the ground or what we call nadir and then flying in that grid pattern and what's especially important to get a nice accurate high-resolution data sets out of this is making sure you have enough overlap so this includes forward overlap as well as side overlap when you move into the inspection and the 3d product workflows your flight plan looks a little bit different so if you're setting out to inspect an object of interest or to create a three model of a building flying in that nadir pattern just isn't going to do it because you're not going to be able to see the sides of the buildings or the side of the object that you're looking to collect so in this case we recommend flying what's called an orbit pattern and that's turning your camera at an oblique angle flying around the object of interest and doing it at a couple different altitudes okay and what this is going to do it's going to allow you to get all the facades of a building or all of the detail of that object of interest to be able to inspect it or create your 3d models so I've been getting asked a lot of questions both today and over the course of the beta is what drones just drone to map work with so drone to map is drone agnostic so what that means is if you have a drone or you have flight management software that's going to go out and control your drone help you automate some of these flight patterns as long as you're getting geo-located imagery we can work with it and so I'll show you some examples here in a little bit so again we built drone to map with the GIS user in mind so what does this mean this means that you don't have to go hunting around for processing options you don't have to know what camera or what sensor your imagery came from all you have to know is that you're going to point to a folder full of images and you're going to hit start and drone a map will figure the rest out for you so very simply you add located images to drone them out they display on the map the photogrammetric processing engine then goes through each individual image and identifies what we call key points so these key points are unique locations in each of the images in this case you can see there's a little grass pattern between some concrete concrete pads and it picks out tens of thousands of these key points over over all of your images and key points that match are then called tie points and that's how the imagery knows how to stitch all of those individual scenes together to get you ortho mosaics digital surface models and other GIS ready products so one of the things we also wanted to do in drone to map was make it really easy for you to share the results so we didn't want to have that requirement to say okay now I'm going to bring my ortho mosaic into desktop and I'm going to click ten more buttons to try to get it out to my users we've extremely streamlined this part of the workflow for you so 2d products can be shared out of tile layers and so you're simply launching the tile layer interface you're selecting which products that you want to share such as a Northam aasaiya dsm hillshade if you had a multispectral camera you produce an NDVI that would be available here as well and then you choose which folders within your org that you want to share the data to and then optionally who you want to share it with whether you want to make it public or you want to share it just within your organization or even specific groups and what drone the map does is since it produced the ortho mosaic it knows the resolution of the imagery it knows what level of detail that it can be cached to and so there's you don't have to worry about how do I set up my cache schema it's automatically going to do that for you and in a lot of cases it's going to give you more detail than what the Azeri imagery base maps are showing so you simply point it to the files you hit go it caches everything for you and it uploads a tile package to ArcGIS online it unpacks that tile package and then stores it as a hosted tile layer inside of RTS online and then from there it's available to your organization interior users to use as an operational layer or even bring it in as a high-resolution base map so in addition to sharing out 2d products you can also share out 3d products so new at the 1.0 release is a new data type that drona map can produce and this is called the scene layer package the scene layer package is simply you can think of it as a compressed textured mesh so those of you familiar with other drone processing software's out there you can produce an obj and FBX these are all different types of textured meshes well at the 1.0 release we're now producing a seam layer package and you do it the same way you do as a tile layer you simply point to your textured mesh and you hit OK this uploads it to ArcGIS online and then it publishes it as a hosted seam layer and it works the same way as the tile layer does you simply drag and drop it into a map and you now have 3d in the web with no plugin alright so in addition to arcgis online dhrona map also works with lots of other apps within the ArcGIS platform first one I'd like to point out is collector so when you're out flying your drones most of these drones have onboard GPS but the accuracy can sometime not be so great so pairing drone the map with collector for ArcGIS you can actually collect GPS points in the field and import those into jono map as ground control this allows you to guarantee the accuracy of your ortho mosaic both horizontally and vertically all the datasets that are produced from drone to maps such as the point cloud the ortho mosaics at DSM s can all be brought into pro or other desktop applications from arcgis to do some advanced analysis or visualization for those who are using operations dashboard you know it's critical to have the most up-to-date information and a lot of times the most up-to-date base map possible in there well Jona map produces these base maps really really quickly so instead of spending days trying to track down to imagery in a matter of minutes or hours do you have a new base map for your operate for your operations dashboard for those who are looking to serve out say the scientific and DDI that comes out of Drona map you can put that into our TS server serve that out as a dynamic image service put some on-the-fly raster functions on it and have that flexibility be available via a service and last but not least we have ArcGIS Earth at the one point to release of ArcGIS earth which happened just last Friday you can now drag and drop those scene packages that drone a map produces right into RTS earth and I'll show that here in a little bit so as I mentioned drone a map has been in beta we just ended a four-month long beta and we worked with I want to say about 2500 beta users over the course of those four months and those who participate in the beta who are in the audience today thank you very much you guys provided a lot of great feedback and we actually implemented a lot of those features and functionality so I wanted to share a few of the what's new since beta so the first and and probably one of the most exciting ones is offline use so you can now check out a license of drone the map and you can bring it into the field with you in addition to the license you can also download portions of your base map so you know if you're going to go collect two or three areas over the course of the day you can actually chunk out those individual sections of the base map both the tiles as well as the elevation and you can bring those into the field with you and have a similar connected experience in a disconnected environment we've added a couple more templates so we wanted to make it extremely easy to produce the products that you wanted to produce without having to change options so we've added additional processing templates in the create new project workflow as I mentioned when I was talking about Collector we've added support for ground control so you can import ground control that have been collected in the field this includes from mapping grade GPSS survey grade GPS and you can really get that survey great accuracy inside a drone to map with these GCPs for those of who don't who don't own a GPS collector and you don't really understand the workflow of going out and collecting good ground control we added a feature called ground control points from the map meaning that you aren't too concerned about the overall accuracy being one centimeter or two centimeters instead you just want it to align to a base map so we're going to let you go and actually create GCPs inside of drone to map by clicking identifiable features in the map and then assigning those to individual images we also made processing more flexible so for those advanced users in the audience today we wanted to give more fine-grained control for how these products were produced so you'll see more options available now inside of processing options we also added a 3d map view and other 3d mapping capabilities to drone the map so instead of just a 2d view you now have a 3d view inside of drone to map so after your 3d projects are done processing you can visualize the results right away before you share them out to your users and last but not least we added process by area so for those who are out flying drones today we know how important it is to have overlap and to get that overlap you have to fly outside of your project area meaning you're probably going to collect some imagery that you don't want to deliver back to your users so by simply providing a clip area you can limit all of the processing and all the results to a specific area of interest so again that can be imported from a file or you can simply just draw on the map okay so let's jump over here to drone to map so when you open drone to map you get presented the create new project page and so as I mentioned when you start drone the map you are presented with these templates and these templates kind of control what products that you're going to produce now there is flexibility you can go in and change it after a project that's created but we wanted to make it really really simple so while you're in the field and you're trying to verify that a collection was successful that all the imagery looks good and you just want to do a quick stitch together you can use the rapid template so it's going to produce a low resolution ortho mosaic a low resolution digital surface model when you're back in the office and you're ready to create those high-res products you can click the 2d mapping and again this is still 2d mapping so it's going to be focusing on the ortho mosaics the digital surface models if you have a multispectral camera it'll produce an end DVI as well building on the 2d datasets is the 3d mapping and this is where you can start to build 3d point clouds that come out as last files three textured meshes that can be shared up to our chests online as well as 3d PDFs that you can quickly share a 3d model with somebody via email and last but not least is the inspection workflow and again this is working with oblique images not necessarily creating map products that get overlaid onto a map but instead working with them in its native resolution inside of a unique viewer alright so let's go ahead and let's create a new rapid project and so I'm just going to give this a name you'll see here that the default project goes to a folder in your Documents you can simply click browse and decide to store that someplace else so I'm going to go ahead and I'm going to click add images I'm going to go to my data folder and I'm going to go into my images here so you can see here we have a handful of individual images now what's nice about drawing the map is that you can work with drone to map imagery even if the imagery does not have metadata embedded inside of the drone imagery so you can see drone a map automatically read these files and realize that the GPS metadata was not embedded inside of the imagery not a problem a lot of these drones will write this out to a log file or to a text file or a CSV and you can simply import that into drone to map alright so I'm going to go ahead and I'm going to browse out here and I'm going to select my geolocation file and just give that a second to load and so now we can see here we have a checkbox for if our data contains headers what type of delimiter we have and we can also map our fields so in the beta version you had to have your GPS data a certain way and if you didn't have a certain way you didn't like it wouldn't let you proceed we made it a lot more flexible now to where you can actually assign the individual fields all right I'm going to click OK so you can see here now I have all my images inside of my inside of my grid here with a green checkbox that means these are able to be geo located and I'll click OK and so dhrona map automatically geolocate all of the imagery and I'm just going to change my base map here to topo so you can see a little bit better so I have about 15 images here these images are over a small Peninsula in Africa and so you can see as I had my imagery base map up not really great here so a little bit harder to do my ground control here what I wanted to do is just kind of show you the speed of drone to map so I'm just going to hit start you can see I did nothing but just point it to a folder full of images optionally I put in my GPS data that this required and you can see down here at the bottom Jona map is automatically going through it's computing those key points it's going to calibrate my camera it's going to do the matches and it's going to go through and do everything in an automated way I don't have to touch anything else so this just takes a couple minutes to run so while this is running let's just kind of walk through the rest of the interface so I already showed you this but you'll see here that I have this palette of online base maps if I'm going out and working in the field I'll also have a section of offline base maps so if anybody saw the plenary we showed the LA Coliseum so we actually took our laptops out into the field as we were mapping the LA Coliseum and we had our base map available here and we did all of the processing in the field to make sure that all the data was good we also have this add data capability I mean if you have a network problem here the add data capability allows you to go into your portal and this can either be ArcGIS online or it can be 10 for one portal for ArcGIS and allows you to bring in different layers from from your portal or artists' online to add to the map we also have our image properties so our image properties else allows us to see what camera was used all of our geospatial information as well as our coordinate system of our source information optionally I could disable any image that I didn't want to include in processing just by unchecking one of these boxes here so a lot of times when you're going out you're flying sometimes as the drones taken off the camera will actually start triggering and you get about ten images that are about four feet off the ground and really no good so you if you didn't catch those as you're loading your project you could identify those and turn those off here all right we also have some processing options here so as I mentioned we go through in the side of the template and we set these processing options for you but before you hit start and you start processing you can come in here and you can change any of these that you want alright I'll come back to that so this ran in about two minutes here so you can see right now it's preparing our data for visualization and you can see here it's taken those 15 images and it's in it and it's added them to our map so I can click on any of these blue points here and I can actually preview any of the source imagery so you can see this is actually pretty detailed imagery now we ran this on rapid so that means that we're not going to get really high resolution so let's go ahead and let's zoom in a little bit and you can see as i zoom in here you'll start to see things be a little bit pixelated still pretty good though right I can make out cars you can make out buildings I've actually ran this entire project before and so let me open up a project that was run this one took about eight hours we let it run overnight and I'm going to go ahead and I'm going to turn off my flight path in my images here and now let's just go ahead and zoom in and let's take a look at the resolution here so you can see this one was run using the 2d mapping template you can see it's much higher resolution okay so as I mentioned we can take now and we can share this data pretty easily so I've run through I've created my ortho mosaic I've created my my digital surface model let me show you show you that real quick so you can see here I can now visualize all the buildings in my digital surface model and so now I want to share these datasets so as I mentioned it's as simple as just opening the tile layer and I seem to be having a bit of a network issue are you are you getting network mm-hmm yeah do you want my face thank you bear with us one moment here while we get the network re-established no internet okay okay I think I'm back okay so as I mentioned I can click this add tile layer button and so automatically I get presented with the products that are available within my project to share so for example I can click ortho mosaic I can click digital hillshade model and then I can pick a folder and I can pick individual groups and I can share those out so I'll come back to that in a second but before I do I wanted to talk about another project that we recently did and this was a project that we did in conjunction with the Army Corps of Engineers and McKim and Crete and so this particular project we actually went out and did some survey grade ground control in the field so I'm going to turn off some of these products here first and so as you'll see here I'm going to change my base map to imagery so we actually went out and we laid about 20 different ground control points and after we brought in these ground control points we open our manage GCPs window and so as you bring in these ground control points you'll see these added into your table now what you need to do after your ground culture points have been added to a tables you need to assign them to images and so what I can do is I can click on one of these individual ground control points and I can see it get highlighted inside of my map and I can click this links button and as I click the links button I have to go and I have to find that particular ground control point and you can see here that they've all been marked in these different five images and to make it a little bit easier to find we've actually sorted these ground control points from closest to the actual point to your image so you basically start at the top of the list here and you work your way down additionally I could click this add from map and so if I wanted to come in here and I want to add a new ground control point let's say maybe up on the streets here somewhere I can simply click and I can add that to the map and so now you can see I have a new ground control point added here and you'll see it down here this is actually orange until I go in and I mark it and when you mark ground control points inside a drone the map you want to mark it in at least two images but we recommend between three and eight images okay so let's zoom out a little bit here I'm actually going to remove this selected one because we don't need it and so after we got this in and it's important and also the note too is that as you add ground control points we've added support for local coordinate systems as well so by default drone and map will automatically output the source data or excuse me the output data such as your ortho mosaics or digital surface models in your local UTM zone but if you're interested in having it in something else simply provide ground control points and your output imagery will take on whatever coordinate system that you specified inside of your DCPS all right so let's turn off some of these points here let's go take a look at our imagery so I'm going to zoom all the way in here so you can see these ground control points a little bit better so what we actually did was we walked around the beach we laid these painted tiles down on the beach flagged them out with pink flagging so they'd be easier to see and then we took a survey grade GPS point at each one of these and so you can see that the the image that's being shown here lines up with the ground control point almost spot-on and one of the reasons why we were working with McKim and Creed on this in auditioning to help us fly this well as we wanted them to do with accuracy assessment so for this particular data set for both the earth mosaic and the point cloud we were at about one point two inch within error the RMS error meaning that the accuracy both horizontal and vertical was under two inches pretty good for a drone and so one of the interesting parts about this data set you can actually see so what you're seeing here is the the world imagery base map and one of the goals of the project was beach replenishment and so we wanted to fly drones to see how much Beach was actually replenished here so you can see these ships are actually dredging in the yard and they're actually back filling sand into this beach and we turn on our ortho mosaic you can actually see they did quite a good job of getting that the rest of that sand filled in there all right so as I mentioned you can also share this as a tile layer so let's switch over to Arceus online now so so I've actually gone through and I've used my Sharra's tile layer button and now I've shared this out to Archie is online as a hosted tile layer so now as i zoom in here I can get that full resolution imagery and you can see here so we I think believe this was cached down to about level 21 or 22 but you can see because this is hosted inside of our tisana and it's been cached it draws extremely fast okay so let's go back to drone to map now another data set I'd like to show you is the is the inspection workflow so this list you'll see right here is all of the individual projects that I've been running but I can also browse to other projects so I'm going to go to my drone a map projects folder and I have a data set here and you'll notice when I go and I open up a project I have this dot d2m file this is the new file extension for drone to map so those who are used to working in desktop and you're familiar with an MX D file drone to map uses a dot d2m file so I'll go ahead and I'll click that and I'll click open and I'll turn off my manage CCP's window I'll zoom in here a little bit make this easier to see and resize my inspection window so as I mentioned when you're going out and you're flying for 3d project or inspection projects you want to do an orbit so this is what an orbit looks like essentially the drone flew around this building that's in North West Virginia this is the Oatlands historical house and gardens and it basically snapped a picture as it was flying around so I can come down here to the bottom of the carousel and I can just click images and now I'm starting to get a virtual tour or moving around that orbit now I can also click a point if I wanted to see say a different side of this I could click simply click and I can move to the side or I know for this particular data set I actually want to inspect the front of the house because there's been some reports of roof damage in this particular area so now I can zoom all the way in and you can see the detail right so because this is the full resolution imagery has not been processed it simply just been oriented and then displayed inside of my viewer I can now draw on this here and I can say I can flag this for somebody to come back in and take a look at it so I can add an annotation so I could highlight the roof as blue and then I could simply add text to this I could flag this as roof repair needed and not only does it exist inside of my viewer so you can see as I move now from image to image I actually get an icon now on this image denoting that there's annotation here so I can come back to this I can zoom in and then if I wanted to share this out with somebody I could simply export this so now I can export that and I can simply go to my apologies I can go to my desktop and I can open up this image and you can see all of my annotations are now burned into this so this can be dropped into a slide deck it could be shared out through an email but an excellent way to get that information about an image over to an expert to do some analysis on it alright so the last data set I wanted to share is a 3d data set and I wanted to talk about the clip area here so as you're going out and you're flying an orbital pattern it's really hard to limit the amount of horizon inside of your images so let's just click on one of these images here and you'll see drone processing software just they it does not like horizon right it's not drone to map it's pretty much all processing software so there's other processing software's out there let you go through each individual image and crop out the horizon we wanted to keep it simple so again what we added was this project area here so what I did was I actually came in here and I just drew a polygon around the area of interest and all processing that now occurs on this particular data set will be limited to that meaning that all of your horizon has now been cropped out so let's move over to the 3d view now and I'll turn off some of these other layers here and so this particular dataset was flown of a fire station in Rancho Cucamonga California and you'll see here we were able to produce a 3d model just of the area that we designated inside of our clip area and so something else to note that this was not collected with a fancy camera this was actually collected with a GoPro so with a $400 GoPro and a very affordable drone we were actually able to fly this area this flight took less than 10 minutes processing took about an hour under two hours we had a full 3d model so that's also as I mentioned you can take and you can share this as a scene layer so again the same way I come in here I pick the product that I want to share and I hit OK I've already uploaded this to ArcGIS online and so now you can see I've got the same 3d model inside of my web viewer that anybody who's a part of my org can now see and take advantage of additionally if I wanted to share this publicly that's just another checkbox inside of drawing the map now the last thing I wanted to show is RTS earth so here I am inside of RTS earth I'm in that same area that we created that fire station you can see I don't have any data in my talk but one thing I wanted to do was be able to show you how easy it is to take a seed package again this is the new format that drona map is producing and I can drag and drop that into RTS earth and before I show you the fire station I'm going to show you one other data set and this other data set is in Germany so you can see I can add I can actually add different data sets in here so this was another data set that we flew and created you can see it automatically geo references it and scales it correctly now let's add that fire station back and you'll see it will take and zoom us over into the correct area of interest as i zoom out a little bit you can see I have the same view now that I had both inside of drone to map as well as the web viewer all right I'm going to stop there and I'm going to turn it over to Kodi but we'll make sure to save a few minutes at the end for some questions all right thanks Tony well can you switch me on to the screen yes all right well Tony I'm sure many of you are probably here to see what Tony presented so not not to start out with encouraging you to leave but this part may not be as exciting for everybody basically the team I work on we work a lot on workflows to get data into ArcGIS so what I want to talk about from this point is bringing the products that Tony showed you at least some of those products into the ArcGIS platform but if that really doesn't sound interesting to some of you I won't be I won't be hurt if you choose to to exit the room at this point so what I want to talk about I do have two slides in here that would qualify as boilerplate but I'll go through them quickly I want to talk just a little bit about the general capabilities of ArcGIS then just a few moments about project planning but really the bulk of this is managing projects from from a drone and the way I portray this is if you've if you're doing two or three or four drone flights a year you probably don't need much of what I'm going to show but if you're really serious and you're doing hundreds or literally thousands of drone flights then I think ArcGIS has some real value and that's really what I want to talk to you about and then analysis I've only got just a few points on that just because you could go any different direction depending on what your project is about so for those of you who aren't familiar with our whole imagery story ArcGIS has a lot of capabilities for working with all forms of imagery we we have a robust data model that allows you to manage literally unlimited amounts of imagery or at least very very large amounts of imagery and make that imagery accessible to your users whether they're on a local area network or or you're sharing over the Internet second once you've gotten your data into ArcGIS I assume everyone knows we do have a lot of analytical capabilities and again this is going to depend on if you're doing surface analysis or you just want to make base maps or if you're doing line-of-sight or 3d view sheds etc a lot of different analytical tools for extracting information out of your imagery and then of course we've got a number of different opportunities to share the imagery Tony showed sharing directly to ArcGIS online many of the clients I work with have ArcGIS server so they're sharing an enterprise at an enterprise scale and one last note if you haven't heard these terms before I'll be perfectly honest I find these a little bit confusing so I thought it might help to put these on the screen you'll hear ESRI talking about a system of record a system of insight and a system of engagement and so I thought that might be helpful to put these up oh and one other thing if am I getting white lines on the screen yes so I apologize for those I don't know why my graphics card is putting up those white lines but those aren't those aren't really part of my presentation okay so with regard to project planning part of what I want to portray or or encourage is the value of arcgis before you get out in the field and again we've got a lot of capabilities that really can support and improve your project planning and i want to be clear i'm not referring to flight planning so we don't do flight line layout in arcgis but what i'm referring to are bringing in the ancillary data layers one of the key data layers of course would be elevation when you're planning your flight line and planning your spatial resolution this is just a simple graphic that I got from Topcon if the if your height above ground is changing you're going to get different resolution you end up with gaps in your project an ArcGIS online through the world elevation gives you that dial-tone of elevation data that you need to support that project planning another another very good example I don't know if Eric Dexter is in the room but I'll give him credit for this idea as long as we're still operating under line-of-sight rules he said it's very important when he plans a project he has to figure out where he's going to stand so he can maintain line-of-sight with his drone for the maximum amount of time ideally the entire flight so he's using elevation and the view shed analysis to help him decide where in the project to stand or to launch from if if he has that flexibility of course there are other data layers you can get out through ArcGIS online weather wind political boundaries if you're mapping a county or a city that sort of thing FAA maps of course are accessible through ArcGIS online and then this last one I give credit to some colleagues from alt avian they've said one of the key things that they're doing in terms of project planning is enabling clients to feed their GIS content directly in to the project planning process and again that may be obvious but but for those of you who either have ArcGIS already or you're serving clients who do I think it's very important to streamline that process of project planning using that client data so I want to talk now about bringing the data products from drone to map into ArcGIS and at this point I want to I'm really focusing on primarily the 2d inspection praat are the 2d products the ortho rectified mosaic and then the digital surface model we're also well these can be thought of as GIS ready because they're they're ortho rectified there they're already going to come into your map and of course the 3d products do as well but we just haven't had time on my team to work and develop some of these workflows for the 3d products so just a matter of timing and then of course the smart inspection photos as well and again this this component of my my architecture is not as well developed but we do have some capabilities with regard to managing those inspection photos that Tony was showing and again within drone to map you Tony showed some capabilities frankly I didn't even know we're there so there's a lot of capability within drone to map but we're really not portraying it as a deep analytical tool really we want you to bring that data into arcgis to to do any any of the detailed analysis so it's going to be important to bring in those oriented photos that tony was showing as well so transitioning to data management i want to talk a little bit about the mosaic data set and that's my second boilerplate slide then you'll know those are gone i want to talk about a model an architectural model that we're recommending for managing large collections that we refer to as the source and derived model and if any of you've heard me speak before you've probably heard some of this i also want to show you and provide some capabilities in terms of automating this because as you may know with the mosaic data set there are a lot of properties there's a lot of power but there's also the potential to check the wrong box and then not know what to do to fix it so our automation tools help show you our best practices what we recommend for configuring the mosaic data sets and then with regard to sharing of course again Tony showed sharing through ArcGIS online but if you've got ArcGIS server and you're sharing to your enterprise with server there are some differences okay so here's my second boilerplate slide if you're not familiar with the mosaic data set it's it's a geo database model that images access to your imagery so we're not bringing images into the geo database it's only a table that references your imagery in the mosaic or within the mosaic data set it allows you to ingest and define custom metadata and again that is in the geo database and you can also apply define processing to be applied so the simplest example of that processing again many of you may have heard this but if you've got elevation data and you want a hill shade or slope or aspect in the old days you used to have to run a geoprocessing tool and let it crunch for minutes or hours and create a new data set and now you had more data to manage with the mosaic data set we can do things like the hill shade instantaneously on-the-fly so if you haven't heard that that's about all I'm going to say about it but if you'd like to know more about those on-the-fly functions let let me know and we can talk later the other component is the dynamic mosaicing and for the ortho mosaics coming out of drone to map that's not as that's not really applicable but again if you're doing multiple projects and you have to break a project into multiple smaller areas then the mosaic data set can do that mosaicing for you and then the on-the-fly processing I mentioned so once this is in the geo database now we can search and sort for data based on attributes such as date such as camera if you find you've got a problem with one of your cameras and you want to go back and find all the projects that were flown with that camera as one example that would be where we could access the data through the mosaic data set okay so the structure or architecture that I recommend in most cases if you're dealing with large collections is to take each project and put it into its own mosaic data set and do QC on that as a collection before you move on so in this case we'd be ingesting the digital surface model and the ortho mosaic and optionally those single frames into separate mosaic data sets so in and if again if you've got some experience with this I admit it a little bit of overkill to take a north mosaic from drone to map and put it into a mosaic data set you're really not adding a great deal of value at that step other than adding key metadata that you may want to get back to that project but I'll admit that's a bit of overkill at this stage but the next step of this overall architecture is then to merge all of those collections into one centralized collection and that way instead of looking for the flight from June or the flight from May or the flight based on the day it was raining that sort of thing you've got everything in one single collection and then the attributes of the mosaic data set will allow you to select those projects you want and of course you can navigate on the map and simply go to Rancho Cucamonga or or go to where the banana plantation was for example and the footnote just so I don't forget to clarify I'm not talking about merging ortho mosaics in with digital surface models so this same diagram would be would apply only to the ortho mosaics you'd replicate that again for the digital surface model data and if you're managing single frames you'd replicate this again and maybe now you can understand why I would encourage that you use some automation so let me go ahead and take you to a quick example of what these automated tools would look like I've already run it for one data set so if I bring in this data I don't know colter here by any chance Pete this is your data if you're in the room so this is some data I got from a colleague at San Diego State and what I've done is brought that ortho mosaic into a Bose mosaic data set and again at this stage there's not a there's not a huge value in bringing that mosaic into the mosaic data set but what I'd like to do now is go ahead and show you these tools and and basically what we've tried to do is simplify the process of managing large collections so in this case what I want to do is specify the output geo database and I've got the demo data set that you may have seen if you've used the if you've used the software and what I I guess I need to give the give a mosaic mosaic to set bass name I'll just call it banana and the drone demamp project folder is basically to load this entire project folder into this tool in this case just for the sake of speed I'm not going to ingest anything other than the the ortho mosaic and now when I run this tool let me scroll back down so I can talk about the geo database when I run this tool what it's going to do is bring the data from that project into separate mosaic data sets and again the key in doing this is really to populate those mosaic data sets with attribute data that's going to be critical to your organization to find that data again and I didn't time this I think it's less than a minute but it it'll take a moment to to come to add this so in this case what I'm doing is adding multiple projects into a single geo database and this is up to your organization how you want to manage it you may you may generate one geo database for March and then another for April you may choose to put every single project into its own geo database and that's just a matter of how your organization works how you're delivering to customers etc to choose how to organize that but then once I've once I've built individual projects and QC those then my next step I think I'm just about done here then my next step is to merge those into this master collection that you see in catalog where I've got all of my digital surface models in one single mosaic I've got all of my ortho rectified mosaics in one single collection and then and again the point of that is just to enable quick access to find those projects of interest on demand I should have timed how long this takes okay it's got to be just about done and thank you most most everyone for staying I'm quite pleased that you're listening to my data management pitch sometimes people within ESRI say nobody wants to hear about data management it's just not that exciting but I'll tell you if you're doing five hundred to a thousand drone projects a year this can become pretty important okay you know what this mate looks like it's taking two minutes instead of one I'm just going to go ahead and cancel it not sure it's terribly important to show you another mosaic in the geo database and actually I'll just skip running the second tool again I don't remember exactly how long it takes but maybe a minute or two I hopefully I've made the point that the idea is to have multiple projects combined into these master collections so I think that was all I wanted to show on that let me get back to the slides because I did have a couple more things I wanted to talk about Tony did talk about sharing and showed sharing through ArcGIS online actually I failed to put ArcGIS Earth on here so we continue to add more options but again if you are working with multiple projects again that magic number of a thousand drone projects a year you might not want a thousand different data sets in ArcGIS online you might prefer to have one single image service that has everything in it and the mosaic data set does support that and sometimes I'm guilty of throwing out technical details and people tell me I shouldn't do this so I labeled it a distracting tangent but if any of you are really working with large volumes of data we have a data format called MRF the meta raster format and there are other presentations and I can talk to you about this it's basically a very efficient way for storing and accessing data from Amazon s3 or the azure cloud but once you've if you follow this architecture you can then serve your image your data as a single image service and then access that for example using the data collection by date attribute and use the time slider or in this case what you're seeing is a time selector which is a widget for our web app builder to quickly scroll through available datasets based on date and then of course if you're dealing with multispectral imagery we can using those on-the-fly functions I referred to you can instantaneously look at a color infrared view a natural color view and NDVI etc or any other functions that may be appropriate for your data so moving to analysis am i doing on time Johnny okay we should still have some time for questions so again I assume you're familiar with at least some of the tools within ArcGIS there are spatial tools so you can work with the DSM and do line-of-sight view shed analysis that sort of thing we've got spectral processing tools to do classification things like NDVI one quick footnote I guess if any of you are doing drone flights for agriculture I'd love to talk to you about turning that turning your imagery and making sure you're looking at surface reflectance if we've got time and you're interested I've got an example of NDVI from a drone company where the water is brighter than the ground and that simply isn't right some of the some of the cameras that are being used for color infrared and agriculture really don't have the proper radiometry to give you a good NDVI so again another distracting tangent of course we've got segmentation I put that by itself because it's kind of both spatial and spectral and obviously the tools are going to depend on the kind of content and project analysis you're seeking okay I've got a couple more quick demos but I'm just going to go through this really quickly because I want to make sure there is still time for questions the first is I voluntarily very common within the drone industry nowadays and I've got an image here this is a hill shade of a DSM these are coal stockpiles provided to us and this was flown by a drone provided by Southern Company and they wanted to do a volumetric analysis what I've done is drawn a quick transect to show you the problem in this particular case is this stockpile is not on flat ground so if you try and run a cut and fill you're either going to miss part of the stockpile or you're going to count some of the ground so one of our engineers put together a tool and I'm not going to run it but I thought I might just show you the model very quickly if I can find it basically he just put together a model to estimate the ground surface underneath that stockpile and again I'm not going to run this but if this is an error you have interest then please let me know the other one that I'd like to transition to very quickly is some server-side capabilities and once again if you're doing five projects a year this is probably not of interest but one of the capabilities of ArcGIS is if you want to share analytical tools with clients that are on say a mobile device and they don't have a lot of processing power on their device you can actually run GP tools on server and then share those as geoprocessing services so what I've got loaded are a couple of different dates of stockpiles two different dates from 2014 and as I flicker between these two the stockpiles on the east side you may be able to see changing I've also got a hill shade of the DSM so you can see the differences in those stockpiles so in this case what I wanted to do was just show if if you've got it if you've got users who need to make these kind of measurements out in the field let me see if I can find my proper tools it's actually labeled lidar tools but in this case we're dealing with a photogrammetric point cloud and basically with this tool I can draw a polygon around the stockpile and then ask the server to do the work and this is an honest demo I do have the processing power within arcgis desktop to do this but in this case i'm actually talking to a server in amazon all i've sent it is that polygon it's doing the work and returns to me a polygon if i open the attribute table you'll see the volume estimate within that now I didn't combine those two so this is not sloping ground this is a simple flat ground kind of case but again if you've got people in the field who need to make these kind of measurements you can share those tools as geo processing services and I think that's about it I've got a bunch of resources here the one disclaimer I need to admit to everybody is the tool I showed for automating ingest of drone to map is not yet on this site so I'm pointing you to some of the work that our team has done for man pre-processed orthos elevation data high-res satellite data etc the drone to map workflow is not there today we simply ran out of time but I can get you a beta copy or I can keep you posted as as we get it ready documented and and debugged so I think that's everything oh and if you're interested in the MRF that's at the bottom but we can talk about that separately so let me stop and see if there are questions for either of us it is the point cloud is la s format so yes go ahead yeah so I can take that so so the so the first question was about last files and so we talked about that the the point cloud that does come out of drone a map is la s the second question was can you process multiple flights together and the answer is yes so as you're going through and you're adding images to your project you can add multiple folders or individual images from multiple directories from multiple cameras correct yes so the cost of drone a map is $3,500 per year per user so this is a subscription license it comes with so I should I should caveat that by saying that is the u.s. price if you're interested in the International price please see your distributor but it also comes with a named user inside of ArcGIS online and 500 service credits as well and I don't believe that we're including it in the personal use license yet but you can stop by the drone cage area and we have some folks there from the sales team who can give you more information about that well and you can get a sample a trial version correct so you can get you can get a 60 day free trial from the marketplace today as well all right other questions so the question is questions are we going to give you an estimate of credits used before pushing it out so that is so that is currently not in the one Dido version but ArcGIS online does have a credit calculator so you can go into that credit calculator and for the hosted tile layers it's one point two credits per gig and if you know most of these drone flights drones typically stay up in the air for under an hour by the time you process it you're talking megabytes of data not gigabytes of data so that's one of the reasons why we wanted to include 500 service credits with a drone a map subscription to allow you to be able to push that data up to ArcGIS online and then get that replenish that service credits every single year every time you renew your subscription and let me add to that very quickly I don't have the slide handy but I did a calculation one time not for drone data but if you had aerial photography I did the calculation for the city of Denver I think it's 500 square miles and at six inch resolution that would cost you $1 a month in ArcGIS online so we do charge you but it's really not very much so questions over here so it is is not currently part of the EPN but like you said we do allow trials of it yes in the middle of the attitude of the altitude so the precision of the altitude is going to depend on the drone that you have and so this is purely at the mercy of the GPS that's onboard the drone a lot of drones today more of the higher-end drones will have RTK capability meaning that it's actually streaming GPS correction to the drone itself in which case you can get both the X Y and the altitude a lot more accurate at the time you're flying but with vertical control I mean that's really right that's a good point so another reason why we use ground control point is that you can help offset any errors and the altitude while you are flying meaning that it's not only going to give you that XY accuracy it's going to give you the Z accuracy in your final product and on the project you mentioned did you yeah obviously I was about 1.2 or 1.3 inches vertical vertical okay yeah yes over here that's a little too big a question for a short answer I don't know do you wanna um yeah maybe repeat the question yeah so I think so so the question was was talking about flight planning and planning altitude above ground and things like that I don't think we're that were the two experts on that but if you if you stop if you stop by the drone zone we've got a lot of our partners who are the drone manufacturers and they've got all their flight management software down there and they can walk you through and great detail on how to do that yeah it's really going to depend on their capabilities yeah and we can probably make recommendations but we're trying to be friendly to everybody so I don't want to do that in a big group yes so so they're really the only time you need a special sensor is when you're doing anything with multispectral right so whenever you want to produce an NDVI image you've got to have some type of IR band being collected as a part of your imagery other than that and the off-the-shelf camera that is able to be Delocated you can use to create any other product so you don't need a special camera to create the 3d products as long as you're getting enough overlap and you're flying it correctly we're going to be able to generate that 3d model from it right okay so so the question was when you're flying a gopro since the gopro camera doesn't have a gps attached to it how do you associate the GPS points with it so this really works differently with with each of the drones let me give you an example so we've been working really closely with the 3dr solo and every time the 3dr solo triggers the shutter of the camera to take a picture it's also triggering the GPS okay and inside of the 3dr solo software called site scan it will automatically match those up for you there's other software applications out there like mission planner that you can bring in a log file and you can bring in images and it'll combine the metadata together using the time stamp okay so as long as with your your your time stamp of your GPS and the time stamp of your image are close enough you'll be able to geolocate all those images that way that's a good question that that's something that we haven't tested so I would say probably the official answer what it will be no that's not officially supported my unofficial answer would be try it and get back to me let me know if it works good answer yes is that Oh an image size so again that's gonna that's going to really depend on what camera you're using what altitude you're flying at there's a lot of other parameters that go into that stop by the drone zone that's what that's all about and those guys are answering questions down there on that alright we must be getting near the end yeah another question so the question is are we looking at taking the DSM and extracting a de M I think the answer in drone to map is probably not within ArcGIS possibly but it won't be next month I mean that's a challenge it's a real challenge sure so we're actually picks for D is actually the processing engine inside of drone to map so we've we've we've partnered with picks for D and we are actually embedding their technology inside of drone to map so you can think of the processing engine behind the scenes and all the accuracy and the experience that that engine brings to the industry is now inside of drone to map just in a much more streamlined GIS friendly manner with an increasing integration over the future releases so one thing if it wasn't clear when Tony talked about taking drone to map out in the field you do not need to take our GIS with you so today they are they are separable you do have to have that ArcGIS online account though which is why you need to check out a license before is that correct before leaving the office okay another question yes so we we actually have two more releases planned this year the first will be a quality release this most likely you'll hear us refer to it as a patch that will come I would say probably in the next six to eight weeks and then the 1.1 version will be aligned with the 10-5 release coming later this year and one of the important things about the 10-5 release and the 1.1 release have drone to map is that you'll be able to license drone to map using portal for arcgis so in the 1.0 release you need to license it through our GIS online in the 1.1 release you'll be able to license through portal for arcgis and be completely behind a firewall yes okay so the question if I understand correctly is terrestrial imagery collected with GPS with enough overlap so again this falls into the category yes the software will do this but we are not officially supporting it during the beta we've had lots of people go out with their iPhones and their Android phones and walk around snapping pictures I did it my backyard with one of my kids little play farms I was actually able to walk around with my phone snapping images and built a pretty nice 3d model the difficult thing with that because it wasn't collected from a drone because there was no standard metadata information associated with it it's really hard for us to kind of stand behind those products right because it wasn't designed for that so again I would say love to have you go out there and test that if that's something that you guys as a user community are seeing and getting it's a feature that you guys want to see in an upcoming release be supported then by all means give us that feedback and we'll try to work it in probably have time for one or one more question before we vacate for the next yes so all of the educational nonprofit licensing is still being worked out I believe that there's more information down at the drone zone right now on that so like I said we've got our sales team down there and you can kind of talk to them about your your specific situation and they can work with you on that go ahead so the question is editing capabilities so we did we did provide the ability to clip products so if you specify an area of interest before you start processing then all of the products will be clipped to that area of interest your point clouds your 3d models your 2d products if you already run a project and for example you wanted to process the entire ortho mosaic and then you want to just clip out a portion of that ortho mosaic you can run the clip after you've already processed but it will only clip your 2d datasets well and it's that's a 2d clip you don't have any Z constraints correct or clipping it directly so right yeah so if you want it if you want to limit the area that you produce for a point cloud or a 3d mesh you want to make sure you provide that area of interest before you start processing and then everything will be restrained after that yes so again that kind of that kind of goes so the question is if you had a camera on a tripod and you were taking pictures of say of this room could you create a point cloud of that that kind of falls into the category of terrestrial imagery again not officially supported but technically doable inside of the product we had thought about including that in the 1 dot or release but we decided not to just to keep things simple in the first release but like I said if this is something that you guys are doing and you guys want to see this in a future release let us know and we'll try to work that in ok yes in the back here so the question is for flight planning can you program a drone to either follow a line or maybe a prior flight yeah that's really a question again for the drone manufacturer and the and the flight planning software but for the most part the answer is yes again that's outside of ESRI and our technology but most of those most of the higher-end or or even moderate drones can do that but but that's not us I think so I think we probably need to clear the room is there another session I think there's probably another session in here but you know we'll be around all week if you want to talk more all right thanks

Channel: Esri Events

Views: 24,471

Rating: 4.9053254 out of 5

Keywords: Esri, ArcGIS, GIS, Esri Events, UC 2016, User Conference 2016, Drone2Map, Drone, ArcGIS Apps

Id: 63qAQJZGab8

Length: 68min 47sec (4127 seconds)

Published: Tue Aug 02 2016