Oregon-based FLIR Systems, a sensor company that makes thermal imaging products for a variety of applications, has made a large strategic investment in CVEDIA, a Singapore-based machine learning and AI startup. Unless you're sensor geek, that may seem like niche news. But the move is indicative of a new reality in an increasingly competitive sensor market and amid a proliferation of AI and machine learning technologies: Great hardware on its own is no longer enough for smaller companies, but even startups can thrive against major players if they pick a lane and package their hardware with smart AI engines.
FLIR makes thermal imaging sensors for enterprise applications, such as field inspection and firefighting. Its sensors come in a variety of packages, including tablets and as smartphone add-ons. More recently, the company has been making the pitch that autonomous vehicles that run on LiDAR and visible spectrum cameras have blind spots that thermal cameras can mitigate.
Also: Dossier: The leaders in self-driving cars
That presents a marketing challenge. Convincing autonomous vehicle developers to use thermal cameras on self-driving cars is challenging enough, but how can FLIR also make the case that its product should be the thermal sensor of choice? There are other thermal sensing companies in the ecosystem already, including competitor AdaSky.
FLIR's answer is to invest in a second company, CVEDIA, in order to differentiate its sensors with plug-and-play AI and machine learning capabilities related to autonomous driving, such as object recognition. FLIR is betting if it can take that development burden off self-driving vehicle companies while ensuring its sensors are maximally adaptable to various self-driving architectures, it will become the most attractive thermal sensor provider for autonomous applications.
That's a departure from the traditional model used by sensor companies, which is to produce task-agnostic sensors for as many applications as possible, and it's indicative of a tightening sensor market.
We've seen other sensor companies packaging their hardware with application-specific features. A company called Viscando makes 3D vision sensors, which has become a tough market with the entry of Intel's RealSense 3D vision camera, which itself has a built-in AI engine.
But Viscando is able to stay competitive by focusing on a niche industry, traffic sensing for various municipal agencies. It's done that by creating a more refined product package than a simple 3D camera, one that includes AI and machine learning related to traffic sensing applications. That market is plenty big enough for Viscando to scale into, but it's small enough that a company like Intel hasn't specifically targeted it for development.
Also: This scrappy self-driving upstart is about to beat the biggest companies in the world to autonomous grocery delivery
Without the AI package, Viscando wouldn't have a leg up over Intel or any other 3D vision supplier, but with a sector specific stack on top of the hardware, it's able to compete. That tells us there's still ample room in a crowded sensor market for hardware startups, even as sensor prices plummet, but only if those smaller players are willing to specialize and find strategic opportunities to develop smart functionality atop their hardware.
That's what FLIR's strategic investment in CVEDIA is all about.
"This investment in CVEDIA will enhance our ability to innovate sensing solutions that enable our customers to more quickly and accurately make their mission-critical decisions," explains James Cannon, President and CEO of FLIR. "The addition of software algorithms that automatically inform a user or system of critical information is a valuable feature that augments the distinctive and rich data our sensors produce."
In other words, FLIR is now in the business of providing sensors and then some. Other sensor companies would do well to take note.