The term “the Internet of things” refers to a vision of the future where many objects—automobiles, appliances, sensors—will be connected to the Internet.
I hadn’t given this idea much thought until I attended the recent SDR ’10 Technical Conference, hosted by the Wireless Innovation Forum. This is an excellent conference, and the keynote talks I attended were well worth my time. What caught my attention at this year’s conference was that three keynote speakers—from Verizon, Huawei, and ST Ericsson—all spoke convincingly about the emergence of the Internet of things. In the views of these speakers, the impending widespread deployment of fourth-generation cellular networks will make it practical for numerous devices to be connected wirelessly to the Internet at little cost.
This got me thinking: If we could connect devices to the Internet effortlessly and at little cost, what new capabilities might that enable? What new products? Lately I’ve been focusing much of my attention on embedded vision technology, so it’s perhaps not surprising that the first thing that came to my mind was connecting vision-enabled systems to the Internet.
Of course, there’s nothing new about connecting video cameras to the Internet. For example, when I’m getting ready to drive to the mountains in the winter, I use this excellent local web site to anticipate driving conditions. But connecting intelligent embedded vision systems to the Internet is a very different thing. Instead of sending a bandwidth-hogging video feed across the network, now we’re talking about a system that interprets what it’s seeing, filters events, and sends only information about specific events or conditions of interest to the user.
An interesting product that illustrates what’s possible is the Archerfish Solo video monitoring system. This is a consumer-oriented smart surveillance camera that connects to a WiFi network and sends email when it sees something interesting—where “interesting” is defined by the user, within certain parameters. I wish I’d had one of these a few years ago when my neighborhood was experiencing a rash of mail theft—later determined to be the work of identity thieves.
The intersection of embedded vision (machines that see) and the Internet of things (machines connected to the Internet) looks like fertile ground for innovative system developers. I predict that we’re going to see many products that combine vision capabilities with wireless connectivity, and that some of these will be big successes. These new products will be enabled by sophisticated processing engines, including processors, FPGAs, and ASSPs.
But powerful chips aren’t enough. Developing a vision-based system is a complex undertaking; even more so for a system that wirelessly connects to the Internet. Few companies have the resources or inclination to develop such systems from the ground up. To enable the design of such systems, silicon providers will have to deliver much more than silicon and programming tools. One key will be system-design platforms—proven designs that can be customized and extended, enabling system designers to innovate, without requiring them to reinvent the wheel. Another will be reusable application components, such as optimized implementations of the OpenCV library.
I believe that there are big opportunities on the horizon for those silicon suppliers that field not only the right chips, but also the right infrastructure to enable rapid development of connected embedded vision systems. Which suppliers will step up to the challenge?
Add new comment