When small unmanned aircraft eventually become commodity products, as they very likely will one day, the people at 2d3 Sensing won’t have to worry. They’re content to let the multicopter and small fixed-wing UAS vendors duke it out for market share – and perhaps some ultimately for survival – while they continue to focus on a critical but seemingly underappreciated component of unmanned aircraft systems: the software that analyzes and adds value to the terabytes of data that will be routinely collected by these systems. What they’ve created serving the US military is extremely impressive, largely unknown, and likely to be of great interest to commercial UAS users. You can see a couple of cool videos showing their technology in action here and here.
To believe that the future for 2d3 Sensing is exceptionally bright is to believe in a very basic prediction about the future of the UAS marketplace: that the value of the aircraft will go down over time, while the value of software that allows UAS users to develop actionable insights from enormous UAS-derived datasets will go up. If you share this outlook, which has proven itself to be true in countless other technology businesses, then companies that are best at adding value to UAS-derived data are likely to be among the best performing companies in the entire UAS industry. 2d3’s exclusive focus on the problem of enhancing, managing, analyzing, sharing and generally producing meaningful insight from UAS-gathered data makes it unique among UAS technology vendors.
The company’s roots serving the US military give it considerable experience with UAS big data, experience that is likely to translate well to the commercial market. As 2d3′s CEO, Jon Damushexplains, the data from unmanned military aircraft include asynchronous streams from many different sensors and of widely varying quality and data types. The first problem addressed by 2d3’s technology could appropriately be referred to as a supercharged approach to data hygiene. The data is cleaned, temporally synchronized, and georeference so that it is telling a complete and accurate story about a specific location at a specific time. Using sophisticated data and video processing techniques developed through the company’s proprietary research in the field of computer vision, 2d3’s software is able to greatly improve the quality of the data that is captured. Among the enhancements: shaky video is stabilized, blurry images are focused, contrast is increased, resolution is improved, lens distortion is corrected, the data is all accurately synchronized and georeferenced, and much more. The merged and processed data stream is so superior to the raw video it’s almost hard to believe that that one came from the other. The features of 2d3s software, which includes client software called Tactiview and server software called Catalina, however, extend far beyond enhancing datastreams. Damush is equally excited about the platform’s ability to help users tag, locate, analyze, and share the data, as well as the ability of the software to create and work with this valuable metadata. In addition, the software is platform agnostic, meaning the data can come from any aircraft or sensor. Whether you measure the datastream coming from UAS in hours or terabytes, the amount of data is large and helping users get value from it is sure to be among the biggest challenges faced by the UAS industry.
2d3 has some pivoting to do before it can effectively pursue the commercial UAS market. While the company’s offerings are not restricted by ITAR, Damush acknowledges that the features and capabilities of Tacitview and Catalina are overkill for most commercial UAS users. Along with the excess horsepower comes complexity and pricing that is also not well suited to non-military customers. But these are good problems, and to address them Damush talks about a cloud-based model and pay-as-you-go or subscription-based pricing for access to the company’s platform within the next six months (likely well in advance of any FAA regulations that legitimize a commercial UAS market in the US). What remains unclear is which application or vertical markets the company will target and how the company’s cloud-based offerings might be modified to suit the requirements of specific industries, including precision agriculture, asset surveys, firefighting, and law enforcement. The company is currently conducting market research to answer these critical questions.
2d3 does have competition from various directions. Many of the large defense contractors that are active in the military drone market have analytics solutions that could be re-purposed for the commercial market. Chief among these is Trimble, though Damush questions how quickly these companies can shift gears and go after the commercial UAS data analytics opportunity. Perhaps more significant are SenseFly and PrecisionHawk, both of which are in the unmanned aircraft and analytics business. The main difference between these two start-up companies (both of which show considerable promise despite having taken a different approach than software-only 2d3) and 2d3 is that 2d3 is platform agnostic, while SenseFly and PrecisionHawk’s software is integrated with their own aircraft, requiring the user to buy into their entire ecosystem. For them, the analogy to the razor (aircraft) and blades (data stoarge and analytics) is appropriate; sell the aircraft relatively cheaply and make money on the services. While other start-up competitors may exist, we see the UAS data analytics space as surprisingly uncrowded (something we expect to change), with 2d3 Sensing well positioned to be a market leader chasing a very big opportunity.
Read the original story here.