ADAPTIVE IMAGE ENHNCEMENNT FOR UNDERWATER IMAGES USING IMPROVED GRAY WORLD AND DIFFERENTIAL GRAY-LEVELS HISTOGRAM EQUALIZATION |
Author(s): |
S. Anjana |
Keywords: |
image processing, image enhancement, underwater images, contrast correction, color cast reduction. |
Abstract |
Underwater images are poor in contrast, blurred and often contain color cast due to the scattering, light attenuation in water medium. This in turn makes captured images in low contrast and object recognition a difficult task. In some cases, underwater images are captured in very low illumination such object detection itself becomes a challenging task. Even though artificial light is attached with the underwater vehicles, the movement of vehicles also causes some turbidity in the water, resulting in limited visibility. Hence, for clear view, underwater images must undergo processing. This processing includes image enhancement and color cast reduction as the image suffers from poor contrast, non-uniform lighting, blurring etc. Even though, there are conventional methods available for enhancement like classical histogram equalization, contrast stretching, white balancing etc., they fail due to the unevenness in the illumination, non-uniform contrast and also fails to condense color cast. Hence conventional approach cannot be an ultimate approach for underwater images. Therefore, an adaptive algorithm which can effectively work for any kind of contrast and illumination is needed. Adaptive Gray World (AGW) and Differential Gray level Histogram Equalization (DHE) are proposed that are worked in parallel. AGW mainly concentrate on color cast present in image while DHE progresses the image contrast. The proposed method finally obtains chromaticity component from AGW and intensity component from DHE. Both of these components are combined to form the enhanced images. The outputs are increased in visibility, reduced in color cast and in most cases yield better quantitative scores. |
Other Details |
Paper ID: IJSARTV Published in: Volume : 7, Issue : 9 Publication Date: 9/6/2021 |
Article Preview |
Download Article |