The Retinex theory aims to explain the perception of color in the human visual system. In order to enhance the quality of low-light and dark video images, it is necessary to increase the brightness range, boost the average pixel brightness, enhance contrast, and eliminate additive noise. To simulate insufficient illumination, images with normal illumination were successively subjected to gamma correction, additive Gaussian noise, and impulse noise. Insufficient illumination compensation was implemented using the multiscale Retinex algorithm with color restoration, additive noise suppression was achieved using a Gaussian filter, and impulse noise was eliminated using a median filter. As a result of this theory, effective algorithms have been developed that enhance the local contrast of an image. Among these derived algorithms, the Multi-Scale Retinex is the most efficient. This article discusses the Single-Scale Retinex (SSR) and Multi-Scale Retinex (MSR) algorithms, as well as the restoration of image color after using MSR.