News like PrecisionMapper from PrecisionHawk being available for free dominated the headlines at AUVSI’s XPONENTIAL event this year, but a quote from one of the keynote presentations caught people’s attention in a different way. Intel CEO, Brian Krzanich, told the audience that “data is the new oil,” as he explained why data would fuel the growth of autonomous applications for unmanned systems. He also showed everyone how that would happen with a demonstration of an automated bridge inspection being done by a drone.
The correlation between oil and data is one people have been talking about for many years now, but it’s a concept that needs to be further explored. Just as crude oil needs to be refined to create usable products like gasoline, data needs to be refined to deliver actionable information, especially in the drone industry. It’s simple enough to send a UAV into the sky to gather a vast amount of data, but who’s going to be looking at it? Why are we collecting this data in the first place? How are the results being brought to action?
These are just some of the issues that operators and stakeholders need to be focused on as they look to take advantage of the opportunities that drones are presenting to them. The role that drones will play in this refinement process depends on the ecosystem they’re being utilized in, and the way in which these systems and solutions will develop and grow needs to be considered.
Drone Data and Information Ecosystems
Data is driving the modern world in countless ways. Whether it’s data that is powering ride-share apps or entire social networks, simple and complicated systems are being propelled by data, some of which is or will be gathered by a drone. That fact represents a major development, since it’s an illustration of how the way in which this data is refined, integrated and processed is changing. It’s an evolution that Intel is focused on, all of which comes back to the power of data that Krzanich was focused on in his keynote.
“We’re seeing a transformation right before our eyes,” said Anil Nanduri, General Manager of the UAV segment at Intel. “What’s happening is that connection points with drones and systems like the Internet of Things (IoT) are being enabled, which is generating a lot more data. That allows us talk about a host of devices that have to connect and share information, some of which was previously unavailable. It’s a trend that has driven this explosion of capabilities, and it’s happening right now.”
Drones and their inherent capability to take to the sky and gather gigabytes worth of data in minutes are in part driving this explosion of data. The trend that Nanduri is talking about regarding how this very data is being integrated into much larger systems like the IoT will enable M2M (machine to machine) connections that will in turn create incredible efficiencies. However, the danger associated with feeding irrelevant data through those systems can’t be ignored.
Juan Plaza has seen this issue evolve firsthand. As someone with experience in enterprise GIS via both manned and unmanned approaches, he’s seen stakeholders and even operators struggle with issues relating to how much of this data they should be gathering, especially when the process to refine that data hasn’t been worked out in advance.
“My concern with the hyper abundance of data these days is that sometimes we collect more than we need for the task at hand,” Plaza mentioned. “In the old days where planes and crew were expensive and the product of the flight was just a roll of film, the missions were very specific and there was no time or budget to fly more or cover more area.”
As an example, Plaza mentioned that things like precision requirement need to be understood in advance. That allows operators to fly higher or lower and to optimize their flying time to accommodate the scale required for the job. If someone needs a 1:10,000 map with a precision of 10 cms (3.93”), collecting data as if you are compiling a 1:1,000 with a 1 cm (0.4”) accuracy is not necessary, and ultimately complicates that refinement process. Determining these mission parameters is the best way to avoid overcollecting and overprocessing data.
A glut of raw data is just one of the issues related to effectively transitioning raw data into actionable information though. For some, the core issue relates to being solely focused on drone data.
While many in the industry have and continue to talk about drones ultimately being able to be seen and positioned as just another tool, their ability to function as ecosystems unto themselves is a factor in what this data refinement process might look like. However, it’s important to note that many view this refinement process in a more complete way, and we can stay focused on the news that came out of XPONENTIAL to find another example of that approach.
“Airbus Aerial was also announced at XPONENTIAL, and they’re focused on collecting vast amounts of data at an incredible scale,” said Kay Wackwitz, CEO & Founder at Drone Industry Insights. “They’re not just focused on data from drones flying close to earth, but also from HALE (High Altitude Long Endurance) drones and satellites. They’ve recognized how this information can enable faster decision making, because the precise aerial footage these systems can provide empowers users in all kinds of ways when utilized correctly.”
The information ecosystems that drones are creating or becoming part of is the key factor in terms of what the refinement of data actually looks like for operators. Are these information ecosystems solely using drone data, or are they pulling in information from various places and sources? It’s a consideration that can be a major part of their success, and that distinction is one people on all sides of the issue need to realize.
Drones Becoming or Integrated into Data Refinement Ecosystems?
When Krzanich was talking about data being the new oil he wasn’t specific about data gathered by drones, and it’s clear that Intel looks at the topic in a holistic manner. It’s an important distinction to make, since drones have enabled users to gather information more efficiently and effectively, but the information they’re getting would often be gathered by some other means if a drone were not available.
Data processing systems that have been utilized for years or even decades haven’t suddenly disappeared, so the question becomes one around how drone solutions will replace or interact with those systems. Are drones going to disrupt all of those solutions or are they going to end up creating connections for these existing systems that are out there?
“Machine learning and artificial intelligence systems don’t differentiate between data sources,” said David Boardman, CEO of Stockpile Reports, which is an image-based stockpile inventory management system. “Drones are really just one more mechanism for feeding data into a solution that’s designed to extract useful information from unstructured data. That said, we’re getting data we didn’t have access to before. They’re giving us a different type of data to throw into the mesh of existing technologies and solutions, but they’re ultimately just one more connection to pump data into that refinery.”
The flip side to this kind of integration plays into what it means to take advantage of a complete drone solution, and it’s a concept that has become very popular of late. Whether users are directly integrating drones into a process they already have or they want to get their actionable information from their complete drone solution, working through what that refinement process looks like is a critical step in seeing success with either approach.
“This discussion about data isn’t a philosophical one, since it’s all being done today,” Wackwitz mentioned. “The main problem that I see is more of an infrastructure problem. It all depends on the infrastructure. Industries really have to catalyze this development by requiring the infrastructure to move data quickly from the drone to wherever it needs to go. It needs to be distilled down to the actions that need to happen based on that data. Maybe it’s going to the IoT or to another ecosystem, but that infrastructure piece between those systems is the core issue.”
The way in which drones become part of or create these infrastructures are being worked out at the highest levels in the industry, but individual users and operators are dealing with these issues today. It’s why the choices that are being made around how drone data is being gathered, processed and utilized are so important.
Refining Drone Data Today
Project needs such as the level of precision that is required have to guide the decisions that are being made. Where and how data is refined stem from the goals of a project, which means the decisions around solutions and processes have to come from those requirements.
“I believe it’s essential to have a final product in mind before embarking on any project,” said Plaza. “UAVs and their missions are no different. Instead of being blinded by ‘features & functionalities’ of specific technologies, let’s clearly define the goals. If we approach our mission from the point of view of concrete results that come from needs relating to specific measurables, data is just another step in the collection and processing chain.”
If you’re a surveyor in charge of an open pit mine and you’re responsible for calculating the volume of materials removed every week from the site, your objective is very clear. If you are photographing an agricultural field to determine plant health, you are going to use false-color or infrared sensors, so the same logic applies. Those are the kinds of decisions and details that need to be established up front, and are a critical consideration in any refinement process. All of which influences the tools, solutions and even systems that can and will be utilized by professionals in various industries.
“If someone is looking to solve a problem like more efficient grading of earth, data gathered from a drone might just be part of that business solution,” said Boardman. “What’s important in that case is that they’re buying a solution, not just data, and not just a tool. Drone solutions can be business solutions, but it all comes back to what kind of problem they’re actually trying to solve.”
Operators who are consciously or unconsciously focused on raw data often find that the process to refine all of it into an answer isn’t something they’re prepared for, which leaves them with a pile of raw data that isn’t especially useful. Few professionals have the time or energy to sort through raw data, and doing so in many cases represents a less efficient use of their time.
It’s not just the ability to be able to meet a specific goal or answer that’s important though. A solution that is able to incorporate a variety of data sets and perform data analysis can open up brand new opportunities related to efficiency and savings.
“I could easily give you 16GB of high-resolution imagery of a bridge that’s been gathered by a drone, which is the typical size for one flight,” said Nanduri. “But then who’s going to look at all of that? What are you going to do with it? It’s great you don’t have to be hanging off a harness to get this imagery, but parsing that amount of data is a very complex task. An inspector is looking for cracks, not 16GB of info. Additionally, there’s supplementary data that can be supplied when that crack is identified which includes reference data, or imagery that relates to a trend. When you have all of that, the inspector has a very specific, action orientated workflow that is incredibly useful and efficient.”
That usefulness is ultimately the goal for any data refinement process, and drones have the ability to gather data that is essential to existing and newly created processes. Data very well might be the new oil, but it’s the refinement of both that has and will continue to change the world.