Open AccessImage enhancement via texture protection Retinex
Author(s) -
Dong Linlu,
Zhao Liangjun,
Wang Jun
Publication year - 2022
Publication title -
iet image processing
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.401
H-Index - 45
eISSN - 1751-9667
pISSN - 1751-9659
DOI - 10.1049/ipr2.12311
Subject(s) - artificial intelligence , ycbcr , computer vision , rgb color model , image texture , color constancy , computer science , luminance , texture (cosmology) , noise reduction , filter (signal processing) , image (mathematics) , color image , image processing
Images obtained in dim light do not clearly represent the target scene, limiting information transmission on the image carrier. This study proposes a texture‐preserving image enhancement method, i.e. ETPR. The proposed method draws the illumination map of a low light image by Max‐RGB, and then applies the weighted median filter algorithm and Retinex to improve the illumination map further. Then, the enhanced image is changed from RGB mode to YCbCr mode and the texture of the luminance component Y is described. It is achieved by denoising the texture of the image to be enhanced, and then the final enhanced image is obtained by the method of sub‐region image fusion denoising. The effectiveness of ETPR is established by comparing it with existing enhancement technologies using public data and real images.