Shipments of embedded vision devices in the automotive, industrial automation, physical security and business intelligence markets are forecast to exceed 14 million units in 2018, up from almost four million units this year.

Utilizing a combination of embedded systems and computer vision, embedded vision enables devices to use video inputs to better understand their environment, applying logic and decision making to video signals.

The maturity of embedded vision algorithms varies by application market. For instance, while embedded vision technology has been active for some time in markets like physical security and industrial automation, the consumer industry represents more of an emerging opportunity.

However, despite the synergies in algorithm requirements across application markets, there are very few vendors that are active across multiple applications. In some markets, like automotive, the long sales cycles and high qualifying requirements have limited new competition. In others, such as physical security, the fragmented equipment market means that algorithms need to be optimized for a large number of products which can act as a barrier to new entrants.

While the software and hardware vendors in embedded vision are unlikely to move into every application market overnight, developments in the automotive space, in particular, should help spur more accurate and reliable algorithms across the embedded vision industry. This trend, combined with increased awareness in the consumer market for augmented reality and gesture recognition, means that demand for embedded vision devices will grow rapidly in the decade ahead.