
The automotive industry has developed several high sensitivity CFAs, such as RCCB, RCCG and RYYCy, by substituting the primary color filters in the standard Bayer color filter pattern with lighter colors. This has had the side effect of lowering color accuracy, and more importantly, color separation of important traffic features such as traffic lights and lane markers. All high sensitivity automotive CFAs retain the red filter given the importance of red lights and signs. Counter-intuitively, this is a sub-optimal strategy, since the ideal red spectral response, being a difference of two Gaussians, has a large negative lobe that cannot be accurately approximated by a color filter. A more accurate method of capturing red is as a difference of yellow and green, which is analogous to the difference of L and M retinal cones that the human visual system uses. Remarkably, this results in both an improvement of sensitivity and color accuracy instead of trading off one for the other.

New, high-resolution CMOS image sensors for mobile phones have moved beyond the single-binnable Quad Bayer and RGBW-Kodak patterns to the double binnable Hexadeca Bayer pattern featuring 4x4 tiles of like-colored pixels. Pixel binning enables high-speed, low-power readout in low-resolution modes and, more importantly, a reduction of read noise via floating diffusion binning. In this paper, we present the non-intuitive result that Nona and Hexadeca Bayer can be superior to Quad Bayer in demosaicking quality due to degeneracies in the latters spectrum. However, Hexadeca Bayer suffers from the weakness of generating Quad Bayer after one round of binning. We introduce novel double binnable RGBW and LMS CFAs, composed of 2x2 tiles capable of 4:1 floating diffusion binning, that are free from spectral degeneracies and thus demosaic well in full resolution and both binned modes. RGBW offers a 6 dB low-light, read noise-limited, SNR advantage over Quad and Hexadeca Bayer in all resolution modes, while for LMS, the corresponding advantage is 4.2 dB.