Dual-Port CMOS Image Sensor with Regression-Based HDR Flux-to-Digital Conversion and 80ns Rapid-Update Pixel-Wise Exposure Coding

Abstract

Today's best consumer cameras typically use computational imaging techniques to digitally enhance images by means of software post-processing to yield both high fidelity and low cost. A common technique, for example, is to combine multiple shots with different camera settings into one enhanced image that has a high dynamic range (HDR). This post-processing-based approach fails when fast motion and/or rapidly changing illumination are present, as often happens in autonomous driving, drone imaging, and action-camera applications, or when the illumination itself is actively controlled (e.g., for depth sensing). These applications require a much tighter temporal integration of: (1) in-pixel processing, (2) pixel readout, and (3) post-capture enhancement. Aiming to address these needs, a new class of ‘coded’ computational image sensors has emerged, with both fine (i.e., per-pixel) [1–3] and coarse (i.e., per-pixel-cluster) [4] programmable exposure control. Some of these sensors offer spatial exposure control for single-shot HDR imaging [1], [2], but require multiple ADC types and/or a number of pre- and postprocessing steps (e.g., adaptive pixel-wise exposure control, HDR reconstruction, etc.). Other coded sensors support a variety of computational imaging techniques (e.g., robust depth imaging, compressed sensing [3]), but their conventional ADCs do not offer HDR readout. HDR sensors exist that digitize the pixel output during exposure, before it saturates, but offer no coding [5], [6].