Characteristics and Performance of Image Sensor Communication

A complementary metal-oxide semiconductor (CMOS) image sensor can be used for communication, in addition to imaging. This paper investigates noise characteristics and system performance for CMOS image sensor communication. Different signal-dependent and independent noise sources are categorized and comprehensively analyzed in conjunction with sensor measurements. Accordingly, a unified channel model in the pixel-matched case is proposed, and the communication signal-to-noise ratio is defined and demonstrated in different parameter settings. Moreover, the channel capacity with mixed noise and bounded input is derived and is proven to be achievable by a discrete input distribution of finite number of probability mass points. Simulation shows that capacity of more than 7 b/s/Hz per pixel or 8-10 b/s/Hz per block of multiple pixels is achievable.