Anyone who has been involved with the geospatial industry can undeniably say, drones have impacted the way they work. Whether it is photogrammetry used for measuring a stockpile or LiDAR for measuring the surface in vegetated areas; these tools have lowered the barrier to entry for geospatial professionals. Before drones, for these types of projects, there were often two approaches: survey boots on the ground, or remote sensing from manned aircraft. Both still have their place and are widely used, but drones have found a niche that existed somewhere in-between. The products we create with these systems are familiar to geospatial professionals. Orthophotography like the backgrounds in google maps, Contours to display terrain, or CAD drawings to show the parcels of a new subdivision. The challenge for professionals has become selecting the right hardware and software to build these familiar products and learning the new toolset. For some, this has been disruptive, for others, it is a natural progression.
While the impact of drones is undeniable for geospatial professionals, the truth is that we are still approaching old problems in a similar way. Drones give us the opportunity to completely reinvent our tactic. We have an opportunity to view these as an edge device in part of a more complex ecosystem of technologies. In a connected world, we need to learn how we can change our workflows to take full advantage of the tools we now have. Tools like drones, augmented reality, 5G, cloud processing, and machine learning give us the opportunity to rethink our approach. To optimize these geospatial technologies, we need to start connecting the dots to create fully integrate workflows that empower people to work smarter.
Creating meaning from drone data
We need to first start with the types of drone datasets we are talking about. Specifically, unstructured geospatial data. To understand unstructured data, we need to define structured geospatial data. Structured geospatial data has a standardized way to relate it to the real world. This means the user can understand where in the world it is, what measurement units are being used, and at what scale and projection it is representing the location. For example, a map can contain a combination of lines for roads (vectors), elevation information (TINs), and orthophotography for a background map (rasters). All three examples have a well-defined structure that includes all the information you need to understand its location in the world. This allows us to have a common way to relate other information as needed.
Unstructured data on the other hand does not contain these characteristics. A still frame picture by itself is not great for geospatial processing. It lacks a lot of important information such as how the camera lens distorts the objects in the scene (internal orientation) and where and at what angle it was taken (external orientation). With drones, we can add this information but even then, it’s difficult for an algorithm to answer the two most important questions about still imagery which is: what it is and why is it important. To answer these questions, we need to give it more context, a relationship with our structured data.
Connecting unstructured data to GIS
GIS is built to answer what, why, and most importantly, where. We structure our databases to answer these things. A database can easily tell you what something is and provide you with the data that helps you understand why it is important as well as how those things relate to other information. The modern world runs on relational databases. Behind most of the things you do day after day, there is a database that is driving the processes including tracking your internet shopping or booking your vacation. GIS is a database containing geometry. Geometry tells the user “where” something is. This adds a whole new level of analysis. Now, not only can we look at the relationships between the information inside the database, the table, but we can also look at the relationship of where that object is, the geometry. For example, if you needed to know all the distribution poles installed before 1970 within 20 ft of a major road, you can easily do that. That is the power of GIS
So, what does drone imagery have to do with GIS? When we link the two together, we create an opportunity to communicate about something that is both easy for human interpretability (the image) and structured for workflow management and data analysis (GIS). This creates a snapshot in time that can be shared with others. It also allows us to ask complex questions both on the drone and in the office. It allows us to collectively pull our subject matter expertise and remove much of the communication barriers that exist and gives us context for better decision making.
For complex connected workflows to work, this is the first connection required. We need to make data intelligence with drones a two-way process. This is fundamentally different than workflows of the past. In the future analysis and intelligence will be conducted both in the field and in the office. Not only will users gather information from the drone, but they will also get information from it.
The cloud, 5G, and data connectivity
In the past, our data was stored in a building behind a firewall inaccessible from outside the building. These days, its more and more common to have that data accessible with different devices in different locations. This “Cloud” connectivity has opened the ability to use this information in different ways. For example. GIS was always analyzed and manipulated using desktop computers. With cloud connected data accessible through a user portal, that information can now be served up to mobile applications where new information can be created and data is integrated directly into field processes.
To have a connected world, you need to have data connectivity. It is difficult to understand the impact 5G will have. 5G will make our data connectivity up to 10 times faster. That means instead of taking an hour to share lifesaving information from the field, it can be done in just over 5 minutes. Quick connectivity from the field to the office is the key to make these workflows possible.
Just as important as getting raw information back to the cloud, it’s important that we can get answers back to the field crews. Context can be the differentiator between a decision that saves millions and a life-threatening choice. For example, if we can get field collected information about an active fire, we can use cloud processing to analyze real-time information to determine the most likely locations that the fire will spread next and we can give that information back to the field personnel in minutes so that they can make informed decisions to take action.
Making it easier for humans
At the end of the day, the technology needs to serve one purpose and that is to make things easier for humans. Computers and algorithms can do some things very well, but we should not underestimate the processing power of the human brain and a lifetime of experience. We believe that technology should not be designed to replace humans. It needs to be designed to empower them to do their jobs with greater ease, efficiency, and information. So how do you build a geospatial data management and inspection system to do that? We believe the answer is Augmented Reality.
Augmented Reality (AR) is the most intuitive way to serve up mapping information. Instead of having to correlate what you are seeing on the live video with your mapping. Your mapping shows up exactly where it should on the live video. This is an example of making it easier for humans. They can now spend less time thinking about what and where something is and more time thinking about what needs to be done. Adoptability is the most difficult task with technology. We want to remove this barrier by making it intuitive and easy to use. This is key in a connected workflow.
Robots working for you
A big step towards using drones for intelligence and analysis is data analytics. Data analytics from drones are only as good as the answers they provide and a person’s ability to act on those answers. One of the major challenges with Artificial Intelligence (AI) from drones is how to both integrate the drone into the workflow but also how to create a pipeline and process of using the information created from them?
The truth is that computers are a long way away from the processing power of the human brain and most likely will never get there. What is a much more likely scenario is using AI to help answer easy questions reliably and quickly to help people make decisions quicker. For example, AI can help sense and avoid to keep the drone from striking an object and falling out of the air, but it’s still not to the point where it can identify suspicious activity while monitoring a location. These types of use cases require both experience and human subjectivity.
However, AI can be tremendously beneficial with drone data. Certain techniques can be used to extract meaningful information from imagery such as a bent utility pole, a water leak, or the size of a crowd in an image. Using AI to give human decision makers information quickly is an extremely powerful tool.
Connecting all the dots
So how do we begin to put all the pieces together and connect the dots? This is our goal at BirdOne. To create meaningful connectivity with these technologies we need to create a human in the loop process that are both simple and usable. The technology should not be the focus for workers, the processes they drive should. At the end of the day its about getting work done.
As these technologies advance, we need to understand the potential of connecting all these pieces together to create tremendous value that improve our capabilities as humans, reduce risk, and increase our efficiency. We are living at a pivotal time where all the pieces are coming together to allow technology to be a tremendous servant of good in the world. We at BirdOne welcome that world where humans are empowered to be an everyday superhero with tools and technology working for you.