Color Image Enhancement Method Based on Bright Channel Filtering

Abstract

The invention discloses a color image enhancement method based on bright channel filtering. The color image enhancement method based on bright channel filtering mainly solves the problems that in the prior art, halo artifacts exist on the edge of an object in an enhancement result image, and time is consumed in the process of treatment. The color image enhancement method based on bright channel filtering comprises the steps of (1) maximizing RGB components of each pixel of a color image to be enhanced to obtain a bright channel image I (x, y), (2) carrying out grayscale closed operation on the bright channel image I (x, y), defining the result of the grayscale closed operation as an I' (x, y), (3) filtering the result of the grayscale closed operation of the bright channel image with cooperation of the bright channel image I (x, y) through rapid connection of a two-sided filter to obtain an illumination component L (x, y), and (4) substituting the illumination component L (x, y) into an illumination-and-reflection imaging model, solving reflection components of RGB color channels of an input image, extending the reflection components of the color channels to the interval [0, 255] to obtain a color image, and using the color image as a final enhancement result. According to the color image enhancement method based on bright channel filtering, only a single input image is needed, a darker image can be enhanced, shadows of the image with uneven illumination can be eliminated, and the color image enhancement method based on bright channel filtering can be used for preprocessing of detection, tracking and identification of targets.

Description

technical field [0001] The invention belongs to the technical field of image processing, and particularly relates to an image enhancement method, which can be used for enhancement of dark images and shadow elimination of images with uneven illumination. Background technique [0002] In the process of image acquisition, due to the change of light source, ambient light, object occlusion and other reasons, the light intensity received by the object in the scene is insufficient and uneven, which seriously degrades the image quality, affects the recognition of image information, and leads to subsequent image failure. Processing such as the extraction of detailed features, target recognition, tracking and understanding cannot be carried out normally. For a long time, the enhancement of insufficiently illuminated and unevenly illuminated images has received extensive attention from the academic community. Braun et al. (Braun, GustavJ., and Mark D. Fairchild."Image lightness rescal...

AI Technical Summary

Problems solved by technology

[0003] Among the above methods, the method based on Retinex theory has outstanding performance in terms of dynamic range compression and color reproduction, but these methods have a common disadvantage: there will be "halo artifacts" at the edge of the scene object in the enhanced result image, that is, after Enhanced image with halo-like shadows in high-contrast edge areas

This is due to the fact that when estimating the illumination, different weights are given mainly according to the distance of the surrounding pixels, and how to maintain the sudden change in pixel brightness is not fully considered, resulting in the distortion of the illumination estimation of the high-contrast edge area.

Images