Low-Light image enhancement algorithm based on HSI color space

Abstract

To improve the quality of low-light image, we proposed a new HSI based enhancement algorithm. This new algorithm enhances the luminance of low-light level images while preserving image contrast and details. First, the original RGB image is converted into HSI color space, then the intensity and saturation components are processed with different enhancement methods, but the hue component remains unchanged, the segmentation exponential enhancement algorithm is applied to saturation component S, then apply the histogram equalization to intensity component I and then the intensity component I is divided into high and low frequency sub-bands with wavelet transform, the Retinex algorithm is applied to the low frequency sub-band to adjust image luminance while the improved fuzzy enhancement is applied to the high frequency sub-band to enhance image details. Finally, reconstruct the component I with inverse wavelet transform, and the reconstructed component I will be synthesized with H and the enhanced S components to get a clear RGB image. By taking advantage of HSI color space and the improved enhancement algorithm, the enhancement of low illumination color image has been achieved. According to the experiment results, this algorithm can obviously improve the visual effect of low light color image.