Improved partial differential equation-based enhancement for underwater images using local-global contrast operators and fuzzy homomorphic processes

This study describes improved partial differential equation (PDE)-based formulations for combined global and local contrast operators for underwater image enhancement. The algorithms remove the limitations of conventional closed-form approaches and the problem of optimal stopping time of earlier PDE-based approaches. Main features include improved simultaneous combination, augmentation and control of various individual local and global processes, guided by optimisation of multiple image metrics. This ensures optimal operation of the algorithms in terms of visual and numerical results and proposed approaches ensure faster and fully automated processing of various images. Additional contributions include the incorporation of a colour space converter, fuzzy homomorphic and post-contrast enhancement processes for certain problematic images. Experimental comparisons indicate that the improved approaches yield better results than several of the proposed works from the literature for underwater image enhancement.