Local Enhancement of SLIC Segmented Underwater Images using Gray World based Algorithm
Rajni Sethi, S. Indu · 2017
Underwater images possess poor contrast and color owing to physical properties of the aquatic environment. Artificial lighting tries to overcome these drawbacks. However it improves appearance of only focussed objects as compared with others. Global enhancement techniques enhance the image as a whole, ignoring the difference in appearance of foreground and background objects. Thus, such algorithms do not work as desired and lead to artifacts like halos, oversaturation etc. This paper presents a novel segmentation based underwater image enhancement method (SUIEM), which first segments image using SLIC superpixel segmentation. The segments are then color corrected using an algorithm based on gray world assumption. Finally, CLAHE (Contrast Limited Adaptive Histogram Equalization), a local version of histogram equalization(HE) is employed to improve the contrast of image locally. Thus, a fast and computationally less intensive local enhancement technique is proposed which performs color correction and contrast enhancement of underwater images. SUIEM produces compelling results as compared with global enhancement techniques present in the literature.