The low-light photographic performance of smartphones could soon get a significant boost, thanks to the development of a new type of color filter. Created by an engineer at the University of Utah, the new filter is said to let in three times more light than conventional filters, resulting in brighter and sharper images with better color reproduction.
Most digital cameras, with exceptions such as the Fujifilm X-Trans and Sigma Foveon cameras, use a Bayer filter to help capture color information. These filters sit over the image sensor and filter light into a mosaic pattern of red, blue and green on a pixel level, before "demosaicing” it into a final image with full color information. However, this absorptive color-filter array is said to be inefficient, as it prevents 50 to 70 percent of light from ever reaching the sensor.
As such, we’ve recently seen developments like Panasonic's Micro Color Splitter which aim to address the problem and get more light to the image sensor. The latest development comes from Computer Engineering professor Rajesh Menon of the University of Utah, who has created a new transparent diffractive-filter array, which lets in three times more light than its Bayer alternatives.
The new transparent filter measures just a micron thick (100 times thinner than a human hair) and consists of a wafer of glass with precisely-designed microscopic ridges etched on one side. This bends the light in certain ways as it passes, and creates a series of at least 25 new codes or color patterns which are, in turn, read by software.
Because three times more light reaches the sensor, and the filter is producing more color information (25 or more codes compared to the traditional red, green or blue) this is said to result in brighter images with more accurate color representation, and virtually no digital grain.
While the filter could be used for any kind of digital camera, Menon is developing it specifically for smartphone cameras where low-light performance is a big issue. He thinks the first commercial products to use this new filter could be out within the next three years. He also sees industrial applications such as for robots, security cameras and drones. For example, this type of filter could allow self-driving cars to better decipher objects on the road at night.
The paper "Ultra-high-sensitivity color imaging via a transparent diffractive-filter array and computational optics" was recently published in the journal Optica.
Source: University of Utah