Illumination Enhancement of Low Light Color Images

Abstract

Image capture under non-uniform light conditions may lose information in the underexposed portion of an image and make the image information invisible to the eye. Due to their poor quality and low visibility, low-illuminated and over-enhanced pictures are not suitable for computer vision algorithms or human observation.The details of the image are critical in many vision-based systems.Over-enhancement hides some features from the image, which might lead to incorrect interpretation of the image’s information and degrade the image’s visual quality. Thus it is essential to control over-enhancement, retain details, and boost the visual quality of an image.We proposed a hybrid technique to increase the visibility of the image and also to provide accurate contrast enhancement. The proposed technique is the combination of the illumination boost algorithm (IBA) and gamma correction. We first process the input image using the logarithmic scaling function and complex exponential function.Then integrate both the images into one image using the logarithmic image processing (LIP) model. Then we use CDFHSD and normalization function to increase the brightness of low illuminated image regions.Finally, the enhanced results are achieved using gamma correction.The experimental findings demonstrate that the proposed method produces best results in detail and contrast enhancement.