Research on an Enhancement Algorithm of Color Low Illuminance Image Based on Bilateral Filtering and Defog Model

Low illuminance image is a kind of common image in dark light condition, usually has high noise, texture loss and other low quality problems. Low illumination image can not be directly used as the input for extracting beneficial information such as feature extraction and image segmentation.The processing of low illumination image usually involves two parts: denoising and enhancement. In the denoising part, this paper firstly preprocesses the L channel of the image, so as to retain more color information in the subsequent processing. Secondly, according to the noise distribution characteristics of low illumination image, the image is denoised by adaptive bilateral filtering. Finally, since the inversion image of the low illumination image called the pseudo fog image is very similar to the fog image. In this paper, the de-fog model based on the image brightness information is used to process the pseudo fog image, and the final low-illumination enhanced image is obtained by reversing the no-fog image. Through experimental simulation, the enhanced image can maintain good texture and color information while de-noising.