Due to the lack of color and “halo artifacts” in enhancement of low illuminance images based on existing algorithms, this paper presents an image enhancement algorithm based on an improved Lab color space Retinex algorithm. The author innovates the image preprocessing, improves the guided filtering considering the deviation of the guided filtering algorithm, and proposes a new method to enhance the filtered image. The overall idea of image enhancement in this paper is as follows. First of all, the low illumination image is preprocessed so that the obvious dark regions in the image are roughly enhanced and adjusted. Then the preprocessed image is converted to Lab color space and the luminance component (L) and chroma components (a, b) are separated out. Using a modified guided filtering algorithm to estimate the incident component luminance value to obtain the reflected component, after the gamma correction is performed, the image brightness contrast is improved by the adaptive logarithm mapping method. Finally, the brightness component and chromaticity component are recombined and converted to RGB color space. Experiments show that the improved algorithm proposed in this paper has a better image enhancement effect and can solve the "halo artifact" problem of the image.

Date Published: 6 May 2019
PDF: 10 pages
Proc. SPIE 11069, Tenth International Conference on Graphics and Image Processing (ICGIP 2018), 110692N (6 May 2019); doi: 10.1117/12.2524449

Published in SPIE Proceedings Vol. 11069:
Tenth International Conference on Graphics and Image Processing (ICGIP 2018)
Chunming Li; Hui Yu; Zhigeng Pan; Yifei Pu, Editor(s)