Automated real-time correction of exposure times in a multispectral camera using an ambient light sensor

Abstract

Maintaining appropriate exposure times is challenging in applications that involve continuous capturing and processing of images in an environment of dynamically changing light. This problem is typically solved by post processing the captured images and adapting the exposure time for the next image based on some brightness metric derived from a previous image. However, this results in a delay between the light changing and the exposure time being properly updated, resulting in under- or overexposed images. In this article we propose using a light sensor to actively measure the changing light and use this to calculate the appropriate exposure time prior to images being captured. We show that by using this method we can capture images that are appropriately exposed for maximum contrast even when the ambient light changes dramatically.