# [Case Study] Autonomous Docked Drone Integration with GIS Systems in Alaska - Event: NestGen ’25 - Published: March 7, 2025 - Length: 29:36 - Watch: https://nestgen.org/library/watch/eMLVMhPZwAg/ > 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. ## Transcript **[0:20]** Hello, hello, hello everyone and welcome to another super exciting session of Nesgjen 2025. Thank you all for joining us. Please say hello and tell us where you're joining from . It means a lot to us. You are now connected to a session that brings together autonomous drones and the power of GIS, all against the stunning and rather cold backdrop of Alaska. I am your host, Bashali , and thanks to my experience and degree in GIS, I have the privilege of moderating this session. In my day-to-day work as a marketing manager at FlytBase, I talk a lot about drone- in-a-box solutions. But today I am here to listen and learn alongside all of you from two brilliant minds who are making these operations a reality. But before we begin, a couple of quick reminders to make sure this session is smooth and fun for everyone. If you have any questions, we'd love to hear them. Leave it in the chat or in the questions and answers section and we will answer it at the end of the **[1:22]** session. Technical problems happen. So if you get stuck like an unlucky drone in a snowstorm, just refresh and reconnect. And if we freeze, wait. We'll be back faster than it takes the drone to return to its base. There will also be polls during the session. So we ask you all to participate. And finally, remember to capture your favorite moments from the session and share them on Twitter and LinkedIn using the hashtag Nesgjen. I would also like to thank our sponsors and media partners for this event. And with that, let me introduce our speakers for today. Our first speaker has extensive experience in the remote sensing and GIS sector, and as a technical product manager for imaging at Ezri. It plays a fundamental role in improving customer workflows related to imaging, lidar, and raster-to-data conversion with RGis . If maps could talk, they'd probably ask Cody for directions. So , welcome, Cody. Thank you. Melanie is with him. Melanie is a GIS analyst at the Alaska Department of **[2:27]** Transportation . Before joining the Alaska Department of Transportation and Border Protection in 2022, Melanie worked as a GIS administrator and analyst in the United States Air Force. Melanie is a true field expert who brings firsthand knowledge of infrastructure monitoring in Alaska. Because in a place where roads can disappear under the snow overnight, you need cutting-edge technology to stay ahead. Thank you again for joining us, Melanie. It's a pleasure to have you at the event. Thank you. That said, I do n't want to take up any more of your time, so I'll hand it over to Cody to start the session. I'll give you the floor, Cody. Very good. Thank you. Let me know when you can see my screen. It should be showing up now. Is our title slide showing? Very good. Thank you. Well, again, I'm Cody Benkelman, imaging product manager at EZRI, and I expect about half of this audience knows who **[3:30]** we are, but the other half may not. So I wanted to show you some slides to introduce myself. Ezre is the world leader in GIS software, location intelligence and mapping, with over 10 million users worldwide. We've grown every year since we started in 1969. Probably before most of this audience was even born. We have provided customer assistance with geographic science and geospatial analysis and what we call the science of wear and tear for many decades. We call our system ArcGIS. I'll come back to our products and capabilities, but for now I just wanted to talk about GIS. If you're not familiar with that term, it refers to a geographic information system that connects data to a map , integrating location data with all sorts of descriptive information that we usually call attributes or, in some cases, metadata. This provides a basis for mapping and analysis that are used in science **[4:33]** and in almost every sector. GIS helps users understand patterns and relationships, and provides a geographic context for decision-making or, in some cases, simply for visualization. When we visualize and analyze data on a map, hidden connections and information emerge. GIS gives us a deeper understanding of our world, which guides decisions and actions. Today we'll talk about drones, the data they collect, and metadata about when and where they flew, who the pilot was, the last 107 certification, etc. This data is usually managed in a GIS. However, if we focus specifically on EZRI and our products and capabilities, we offer a truly comprehensive geospatial platform . ArcGIS is the name of our general system that supports individuals, organizations, and communities, from GIS and IT professionals to mobile and knowledge workers in the field, executives, and the **[5:36]** general public. ArcGIS is designed to create systems and solutions. These systems support almost all types of geospatial data, including imagery, LAR and 3D models, and many related capabilities for mapping, spatial analysis, and much more. Data management is also a core capability, as many of our customers' use cases involve terabytes, if not pabytes, of data that is often stored in the cloud, providing global access. ARGIS is available as installed software, platform as a service, and software as a service. Argis meets all modern criteria for an enterprise-ready technology. It is cloud-based, can be native to public cloud providers, GPU-accelerated with a reliable development lifecycle, supports standards and specifications, and integrates with common IT systems and infrastructure. Therefore, ARGIS supports all forms of **[6:39]** geospatial data. But if we focus more specifically on 3D images and products within EZRI , we refer to this as reality mapping. I'm sure you've heard the term reality capture, and this is Ezre's interpretation. Reality capture with emphasis on placing all data products in their precise geospatial locations. Now, within Argis, our reality mapping capabilities can process drone images or images from large-format aerial cameras or satellites. And we can generate all the data products you would expect, including surface models, real orthoses, point clouds, and high-quality 3D meshes. Additionally, we have capabilities shown in the bottom left that are somewhat unique , including what we call oriented images, which is the ability to work with non- cartographic images such as iPhone photos, 360° images, and geospatially tagged video within the context **[7:41]** of your GIS. Now, following the creation of these different data products, our users can also apply a variety of analytical tools, and we offer flexible options for hosting and advanced visualizations, whether behind your firewall or via the web. Just one last slide to mention our drone-specific products. We have a free iPad app called ArcGIS Flight for planning and controlling drone flights based on your own GIS data. Then we have desktop processing options called drone to map and an alternative for pure cloud processing called vision scanning. Now, if any of this interests you, I'll be available to answer questions today, and my contact information will be on the last slide. But enough about background information. I want to reserve the rest of this presentation for Melanie to tell you about the important work her team has done and continues to do in Alaska. I'll give you the floor, Melanie. Thank you, Katie. Hello everyone. My name is Melanie Wilson and I am **[8:46]** a GIS analyst at the Alaska State Department of Transportation, in state aviation. I run our GIS program and help manage our UAS program at our remote sensing lab, and we are headquartered in Anchorage, Alaska. I'm here to talk about the comprehensive automation of our image collection using drones, satellites, and robotic vehicle generation. We are taking drone data and remote drone data and automating it on the state network in our processing engines, and making the data available as image services and video services for our department. We are very excited about our reality mapping and GPU acceleration, which allow us to make data more accessible, discoverable, and efficient within our department. We are also beginning to standardize the way we go out, collect data individually, process it, and link it to a working model. Well, this is our **[9:52]** vision for the future with our systems and we are getting very close to the complete delivery of digital products, from the studio to the build, the images and the product lifecycles. Okay, great. Next slide. As many of you know, we have more sources of remote sensing images and data than ever before. In fact, 90% of the data created was created in the last two years, which is incredible. So with all these new data products, like the LAR structure for moving satellite imagery, etc., we're trying to figure out what we're doing with these products and where we're going to go from here. Next slide. So far, our drone program has created more than 30 terabytes of raw data. And that's without including the processed data. And it's crucial to find out where they are stored. **[10:57]** How do we process this data? And after several years of drone operations, we are finally starting to figure out how to scale our drone operations and manage these huge amounts of data. So collecting images with drones has become much more reliable and we can have more routine operations, scheduled operations, especially using the drone in a box. So we're looking at how we're going to collect data for avalanche footage, runways at our airports, critical infrastructure, and how we're going to share these data products more efficiently and remotely with our department. Next slide. Okay. So we are seeing this digital transformation model on our business platform. And, so far, we **[11:58]** can take reality mapping, our IT sensors, forecasts and simulations, and analytics, and AI machine learning . And our ultimate goal is to turn it into a more autonomous , adaptable, and generative model with our operations so that we can take all this data that we collect and then say, "Take what we specifically want," and then use our analytics to answer questions we have about our environment in Alaska . Next slide. Well, none of this was possible two years ago. Much of this technology is emerging, which is incredible. Today I'm going to highlight how we're going to bring all these technologies together. Using drones and drone docks to operate in remote Alaska because, as you can imagine, Alaska is a huge area and it **[13:00]** 's not feasible for us to be in every location to collect these images. That's why we're deploying drone docks to help us with data collection. So, whether it's an emergency response like a landslide or avalanche , we're mapping our airport environments; drones have made it possible to collect real-time data and automate the data workflow to our office in Anchorage . Next slide. We have partnered with Skyo and are currently using their X10 platform and have one of their docs, as you can see on this slide, which we are experimenting with. We are excited to begin building our fleet of drones and docks to support remote operations in Alaska. We currently have 22 X10 drones that we use to capture photogrammetric and thermal data. But what happens on our back-end is really the magic part. These **[14:04]** devices in the field can now collect data using Starlink, Wi-Fi, or any type of internet connection, and the data can be uploaded directly into Sight Scan for ARGIS, as Cody mentioned, which is a cloud-based drone mapping software developed by Ezri. Site Scan allows users to plan, capture, process, and analyze drone image data. This allows us to create 2D and 3D results such as orthomosaics , digital terrain models, point clouds, and all within a comprehensive platform. These data products are used to create dynamic image services that can be shared with our users. Next slide. When we think about remote operations, it really seems like science fiction that we can collect all this data and automate this process from storage, processing, and publishing to image services. But, again, **[15:05]** just to reiterate, this technology didn't even exist two years ago. We are using these systems in remote parts of Alaska, where we have some of the most hostile environments in the world. The remote flight deck allows users to control these systems remotely from anywhere they have access to a computer. We have geofencing systems configured and require 107 licenses from our operators. Next slide. Codia I think slide sharing has stopped . Do you love me? Okay. I'm not sure. We can...You're not... You're not seeing my slide. I've shared it. Yes. This is the next slide. Yes. Great. Forward. Okay. Thanks, Cody. **[16:06]** Okay. As we looked at this, all these third parties that are joining in... Much of this is cloud-based and these systems are no longer locked into a specific infrastructure. So our X10 drone data is automatically sent to the Skyio cloud and to Site Scan via an API. And, uh, we're using APIs and webhooks to connect this data, which allows us to transfer all this data to the cloud, and it's essentially our canvas for deciding what we're going to do, how we're going to manage all of this. Next slide. Last year we worked with the Skyideo team to set up an API that allows us to automate the flow of drone flight data to the Skyideo cloud for on- sight scanning. This was the first data integration that Skidia had set up, which was very exciting. This allowed drone operators to fly a mission and fully automate the data workflow from image **[17:12]** capture to processing in the scan-to-sight, where 2D and 3D products are created. The way this works is that the drone operator enters an access code known as a flight name for a mission, collects the data, and then it is automatically routed through the SkyO cloud and sight scanning. Next slide. So, by scaling up these drone operations, we can program them to become routine and obtain repetitive data collection, whether it's a landslide we want to monitor or an avalanche slope, which opens up many options for using them. Next slide. For these operations, we need FAA authorizations. All our operators have 107 licenses. We need to have a certificate of approval, a COA, and fly in approved airspace. Therefore, **[18:17]** we put a lot of work into making sure we collect this data correctly. Next slide. We are collecting 15 terabytes per year. It is very important that we have data governance to manage our data effectively within our system. Therefore , it is necessary to have metadata standards for images, storage standards and file names , data format standards and file types. What you see in this image is that about a year and a half ago we responded to a flood in the Arctic and the flood completely blocked the road, which was one of the main highways into the Arctic. In 3 days, we deployed a drone and a dock, and were able to collect data while repairing the road. You can see that we connected the drone with **[19:23]** Starlink and were able to see in real time in our Anchorage office how all these construction operations were progressing. It was quite impressive. Next slide. Automating this data workflow is essential with images, videos, LAR, and vector data. Our vision is to create a single, unified, and uniform platform within the DOT in Alaska to integrate 3D data and our GIS to provide an accurate representation of reality. We have our automated image scanning system and then we can offer our images as image services through our company Archip, and our videos go directly to our company and can be shared as web experiences and applications. **[20:24]** Next slide. This is extracting information. This was a landslide that occurred in Wrangle, which is in Southeast Alaska, and it was basically straight onto the highway and affected many properties and there were even several fatalities. To monitor the landslide and see what changes have been detected over time, using our images is key to understanding how this happened and what kind of landslides we might expect in the future. Next slide. And this is on Kodiak Island in Alaska. A remote island. This is just looking at the line work and it's an orthomosaic taken from a 2016 study. You're seeing **[21:28]** utility data overlaid with images, the line work. Well, it's very exciting to see all this in a web-based application . Our users can take the images and overlay all this vector data and then help make informed decisions about where we might want to place utilities in the future. Very useful for our analysis and just for visualization. Next slide. Good. So, you know, spread the word. One of the keys to what we're doing is making sure we can share these images and vector data with our users. And what you're seeing now is also the Wrangle landslide. Taking the images, analyzing the detection of changes over time, comparing it with rainfall and how much the slope instability **[22:35]** changes over time. Then, again, we share that data from our site scanning environment with our image server and with our users so that they can also perform analysis and visualization. Next slide. Finally, we return to this digital transformation model and, currently, where we are is where the star is and we are moving towards the AI ​​ approach. So, in reality, we just take our data, most of which is public and stored in the cloud, and we can use AI as a tool to extract the data in order to answer questions about our environment, our emergency responses, and any data we collect. So that's where we are. Next **[23:37]** slide. And that's all. Thank you so much for joining us, and I welcome any questions you may have. Thank you so much, Cody Melanie, for sharing all these interesting perspectives. It's great to hear about the digital transformation model Alaska has adopted, the stage they are at, and how they are leveraging SkyO for port operations. With that, I already have a question. So, what are the current applications for which dogs are being used in Alaska Duty or the department? Any application where we are primarily using EZRI products. Well, we're using Site Scan, our enterprise RJS image and video server; again, all of that is within the RGS environment. I think that answers the question **[24:38]** about the application. The question was more from the point of view of the use case . What are the different use cases that documentation systems currently support? Okay, great. That makes sense. Well, primarily for emergency response, we've positioned our dock at the base of an avalanche slope in Juno, Alaska, so we can collect repetitive data with drones and see what that avalanche slope is like, you know, what the trends are over time. We also have a drone dock located at the base of a landslide so we can collect repetitive data about that landslide. Well, mainly for emergency response at this time . Understood. There's another question from Justin. I think I'll open it for both Cody and Melanie. Due to the Starlink data upload bottleneck, are you considering local processing of raw data and only updating the results like Automosics? That's a great **[25:39]** question, and it's a timely one because, in fact, we just chose Armada, a company that installs Edge Light data centers. In fact, we just implemented a data center at our Anchorage office. The model will change a little for us now, but we will continue to use Starlink remotely and that data will go directly to our Edge Light data center and then we will continue to use the Azure cloud for long-term data storage. As for EZRI, we certainly have many customers who use different configurations exactly as Justin has suggested in some cases, such as what the Alaska Department of Transportation is doing , since their drones are remote. I think it seems to be working very well for loading individual images, but in other scenarios, people will undoubtedly perform the processing locally and then just update the data **[26:43]** products. Very good. I think we're nearing the end of the session. We have a couple of minutes. So, Cody Melanie, do you have any final thoughts before we close or end the session? Go ahead, Melanie. Oh, no. I appreciate everyone coming and listening to our talk, and I'm excited to see how this technology continues to evolve and how we can use it to our advantage. So yes, I appreciate the opportunity to speak . I'm really impressed with the work Alaska is doing, given the harsh environment. You know, in many drone flights you wait for perfect weather, but in Alaska it's a little harder to just sit back and wait. Very impressive. Yes. I think one question that came up was, what is the biggest challenge of operating these autonomous systems in such extreme environments as Alaska? If you'd like to share some ways you 're taking advantage of **[27:49]** this without considering the limitations of the environment. Clear. Yes. I think wind energy is a big challenge. We learned that we needed both DJI drones and Skyo drones because some of them have different levels of performance. For example, dealing with strong winds. But also with the drone docks, you know, we've had a lot of challenges with the cold and we've been trying to figure out how to keep them warm enough to continue operations and keep them connected to electricity because there aren't many places where there are power lines throughout the state. Therefore, we have to improvise. But yes, thank you again. Cody, Melanie, thank you for sharing this information. I think the way they are automating the workflow, the practical cases, both landslides and avalanches, mainly for post-disaster mapping and information generation, are going to be very critical situations. Therefore, I believe we have a lot to learn from **[28:52]** this session. Thanks again, and uh, yeah, uh, I hope the audience enjoys the event and gets something out of the sessions. And please, stay tuned. We have many more sessions coming up. So you can check the schedule and, uh, yeah, see you at another session. Thank you all. Thank you for your time. Thank you. Yeah.