Color correction for RGB sensors with dual-band filters for in-cabin imaging applications

Abstract

Automotive in-cabin monitoring and surveillance imaging applications operate in the visible band during daytime and in the near infra-red band during nighttime. Utilization of the same camera for imaging in both bands is advantageous as it reduces system complexity. This work evaluates four methods for singlecamera dual-band imaging that vary by combinations between color-filter array, i.e., RGB and RGB-IR, camera filter, i.e., infrared-cut filter and dual-band filter, and processing pipeline, i.e., standard RGB and RGB-IR. We conclude that the optimal combination is RGB sensor, dual-band filter, standard color imaging pipeline, and a color correction matrix that is optimized for the visible and near-infrared bands. This method holds the net advantage of system simplicity and sufficient image quality in applications where high color accuracy is of a secondary priority.