There's a type of camera technology emerging with a view of the world similar to what a honey bee sees. The images appear blurry and hazy, but if you're a bee, good enough for finding flowers and people to sting.

It could also be perfect for the Internet of Things by making it cheap to add vision capability to just about anything.

That's the idea put forth by Rambus, a company that designs technologies and then licenses them, for its lens-less sensor. The sensor captures light and relies on computation to shape the data into an image that's good enough to tell whether someone is in a room or a door has been left open. It can also be used to activate an optical lens if a higher-resolution image is needed.

The key benefits of the lens-less technology are its size, cost and power usage.

"What we think this technology will enable is eyes everywhere," said Patrick Gill, senior research scientist at Rambus. He said the technology can cut the cost of putting low-resolution imaging onto an IoT device "by a factor of 10."

The lens-less sensor can be manufactured using processes now used for other sensors. It is very inexpensive for product makers to add sensors that detect motion, vibrations, temperature, pressure and noise because these microelectromechanical systems (MEMS) are made with the same low-cost wafer fabrication processes used for chip manufacturing. These sensors only cost pennies.

A combination of improved manufacturing techniques and demand is driving prices down, said Randall Restle, director of applications engineering at Digi-Key, a major electronic components distributor.

Today, "you can get a 32-bit microprocessor for around 50 cents apiece in reasonable quantities, like a thousand," said Restle. "[There] has been such a collapse of price."

Camera optics are relatively expensive to make because they take multiple parts, including a lens. The Rambus sensor uses a diffractive phase grating technique that splits the light wavelengths, creating a "jumble of spirals," and then computes the image from this raw data.

The Rambus lens-less sensor is still in advanced development, but moving into development, said Gill. The plan is to have it in open beta or full product release in the next year or two, he said.

The lens-less approach may sacrifice high resolution, but "the animal kingdom has shown that that kind of level of vision can be sufficient to navigate environments," said Gill.

Gartner recently predicted that a smart building could contain more than 500 smart devices by 2022. The estimate invites incredulity until you consider how cheap it will be for a maker of anything to add a sensor, wireless radio and software -- which may be open source -- to just about anything leaving a factory floor.

This device proliferation would be impractical if building owners had to run around replacing batteries. That's why very low power sensing technologies will be critical, along with the development of energy-harvesting systems that can capture energy from motion, such as footsteps, or friction.

Gartner tracks the prices of the sensors, and they continue to fall; the cost of enabling a device may be no more than $1 before long.

The $1 estimate can be seen as a basic cost for adding a network and processor chip, excluding integration costs, said Nick Jones, an analyst at Gartner. He's also assuming that embedded software costs will be "approximately zero" since those are amortized across a large number of devices.

Gartner's prediction doesn't assume that all the protocols and interoperability issues will get sorted out. Jones expects it may take as long as two decades for that to happen.