Enhancing Colour Images of Extremely Low Light Scenes Based on Image Formation Model

Abstract

In this paper, we augment a novel approach for removing motion blur and noise in the images which are captured in extremely low-light conditions. However, motion blur and heavy noise are produced in the captured images of extremely low-light scenes. So, our goal is enhancing the low light colourimage. The algorithm used for enhancing the colourimage is “Image Formation Model (IFM)”. We propose a novel method/ imaging system that can capture RGB image and NIR images with different exposure times. An RGB is captured with long exposure time to get colourinformation and a NIR image is captured with a short exposure time to acquire the structure of a scene. Using this image, we reconstruct a clear colourimage  sequence i.e., a latent colourimage  using an adaptive smoothness condition based on gradient and colourcorrelations.