Image Quality Enhancement by an Image Sensor With a Blockwise Control of Resolution, Frame Rate, and Exposure Time

This study proposes a blockwise-controlled CMOS image sensor (BWC-CIS) that adjusts the resolution, frame rate, and exposure time for each block of pixels. This enables the use of a scene-adaptive imaging method that optimizes imaging conditions based on the characteristics of subjects, such as brightness, pattern detail, and motion speed. The prototype sensor divides a <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$2.6~\mu $</tex-math> </inline-formula>m <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$1024\times 1088$</tex-math> </inline-formula> pixel array into <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$16\times 17$</tex-math> </inline-formula> control blocks. Photodiode (PD) charge transfer control switches are implemented on each pixel. Each pixel shares a floating diffusion (FD) with every four pixels, which facilitates readout pixel selection, pixel binning, and exposure time control. Despite its simple configuration, each block can be assigned one of the four modes: normal mode for standard resolution and frame rate (60 frames/s), fast mode for half resolution and quadruple frame rate (240 frames/s), bright mode for quarter exposure time (1/240 s), and low-light mode for half frame rate (30 frames/s) and double exposure time (1/30 s). By switching between these four modes according to the characteristics of the subject, the image quality can be improved while suppressing the data rate. We propose an adaptive modulation transfer function (MTF) model to assess image quality across different resolutions and exposure levels, and compare MTF characteristics of moving objects between normal and fast imaging modes. Experiments with the prototype sensor demonstrate a dynamic range (DR) expansion of more than 12 dB compared with previous sensors fabricated with the same process technology. In addition, they show image quality enhancements, such as reduced motion blur and prevention of overexposure and underexposure.