Why computers that ‘see’ are a hot technology
What’s a beer company to do if it wants to ensure that the thousands of bottles of booze it produces daily are filled with just the right amount of liquid, contain no unwanted residue, and aren’t broken?
It uses computers to keep an eye open for problems.
Machine vision technology, which lets computers see and understand their surroundings, is undergoing a dramatic boom, according to a report Wednesday by the Association of Advanced Automation (AIA), an industry trade group. Companies are spending an increasing amount of money on the sci-fi technology that helps electronic devices recognize images, much like humans can.
In the first quarter, AIA found that North American sales of machine vision gear like cameras and vision sensors grew to $520 million, up 22% from $426 million during the same period last year. Sales of the components that make up machine vision technology, like lasers and LED lights, reached $84 million for the first quarter this year, up 11% from the $77 million in sales for the same time period in 2014.
The AIA says that first quarter machine vision sales were a record high for the first quarter, explained Alex Shikany, AIA’s director of market analysis. For the full year, Shikany said he expects sales to reach $2.5 billion compared to the $2.28 billion of sales for 2014.
Shikany cited the rapid advancement of machine vision technology as the main reason companies seem to be buying more equipment. Companies can now buy cameras the size of quarters that can capture and process high-quality footage that just three years ago wouldn’t have been possible, he explained.
AIA based its finding on a survey of nearly 80 member companies that range from camera manufactures to software companies working to improve the ability for devices to recognize images.
More companies are interested in embedding the technology into other equipment. Car manufacturers, for example, buy sensors and camera gear to add to their vehicles to help alert drivers to blind spots, Shikany said.
The growth of robotics in factories is also helping to drive adoption. Vision technology has evolved to the point where robots can interpret their surroundings and move freely in a facility. It used to be that coders would have to program a robot to move from one location to another.
At the same time as machine vision has witnessed growth, the field of computer vision—a closely related and overlapping technology—has also been booming. Computer vision basically refers to non-industrial uses of the similar vision technology, such as improving the results of Internet search engines, like Google, to recognize and contextualize images in searches.
“It is all on a continuum,” Shikany said.
In the past it seemed as if only manufactures like Cognex (CGNX) and Allied Vision were interested in the vision technology, but now the big Internet companies are jumping on the bandwagon. Google (GOOG) has been a member of AIA for a couple of years, Shikany said.
Facebook recently made a splash with the advancements the social network has made with facial recognition software that can identify a person even if their face was covered up. It’s the power of Facebook’s software engineers and the algorithms that the company has been developing that have led to its image-recognition breakthrough, Shikany said.
But it wouldn’t surprise Shikany if the hardware Facebook (FB) is using to research its computer vision advancements was bought from traditional machine-vision vendors. After all, software and hardware have to go together in order for the technology to work.
In the case of a beer company, machine vision can do the job of a human worker, without ever getting tired. Imagine watching thousands of bottles pass by, day after day, to catch any that may be duds.
Subscribe to Data Sheet, Fortune’s daily newsletter on the business of technology.
For more on Facebook’s vision tech, check out the following Fortune video: